KR102313656B1 - Method for, device for, and system for analazing brain image - Google Patents

Abstract

A brain image analysis method in accordance with an embodiment of the present application comprises: obtaining a first correction parameter for correcting a first morphological value associated with a first brain element; obtaining a second correction parameter for correcting a second morphological value associated with a second brain element different from the first brain element; acquiring a target brain image; segmenting the target brain image into multiple brain regions including a first region corresponding to the first brain element, a second region corresponding to the second brain element, and a cranial region to obtain a third region associated with the first region, the second region, and an internal region of the cranial region; obtaining a first brain morphological index associated with the first brain element based on voxel data corresponding to the first region, voxel data corresponding to the third region, and the first correction parameter; and obtaining a second brain morphological index associated with the second brain element based on voxel data corresponding to the second region, the voxel data corresponding to the third region, and the second correction parameter. In accordance with the embodiment of the present application, the morphological indexes based on the medical image of an object can be more accurately calculated by appropriately applying the correction parameters in consideration of scan conditions in which the medical image is acquired or a position of the target element.

Description

Brain image analysis method, brain image analysis apparatus, and brain image analysis system {METHOD FOR, DEVICE FOR, AND SYSTEM FOR ANALAZING BRAIN IMAGE}

The present application relates to a brain image analysis method for analyzing a brain image, a brain image analysis apparatus, and a brain image analysis system.

Due to the improvement of image segmentation technology, it is possible to segment a medical image to calculate an auxiliary diagnostic index related to a disease, and thus, the field of medical image analysis has recently been attracting attention.

In particular, medical image analysis technology is being used in many ways to provide an auxiliary indicator for diagnosis of dementia, and it is essential to more accurately calculate the morphological characteristics of a specific area from a medical image to provide objective dementia diagnosis assistance information. .

However, the conventional medical image analysis technology has several limitations due to the presence of errors in the medical image itself, such as artifacts that are not suitable for image segmentation in the medical image. Accordingly, there is a limitation that diagnostic auxiliary information for the same object may be different. In addition, the conventional medical image analysis technology uses a method of performing image segmentation after standardizing a brain included in a medical image with respect to a standard brain model to provide completely “personalized” diagnostic auxiliary information for an object. It has the problem of not being able to do it, so it has limitations in situations where high accuracy is required.

Accordingly, there is a need to develop an image analysis method and apparatus capable of providing identifiable diagnostic information to a user even when noise is included in the image. There is a need to develop an image analysis method and apparatus for obtaining diagnostic auxiliary information in consideration of a scan condition in which a medical image is captured and for obtaining personalized diagnostic auxiliary information for an object.

An object of the present invention is to provide a brain image analysis method, a brain image analysis apparatus, and a brain image analysis system that provide information related to a brain image.

The problem to be solved by the present invention is not limited to the above-mentioned problems, and the problems not mentioned will be clearly understood by those of ordinary skill in the art to which the present invention belongs from the present specification and the accompanying drawings. .

The brain image analysis method disclosed in the present application includes: obtaining a first correction parameter for correcting a first morphological value related to a first brain element; obtaining a second correction parameter for correcting a second morphological value associated with a second brain element different from the first brain element; acquiring a target brain image; The first region by segmenting the target brain image into a plurality of brain regions including a first region corresponding to the first brain element, a second region corresponding to the second brain element, and a skull region; obtaining a third region associated with the second region and an inner region of the cranial region; acquiring a first brain morphological index related to the first brain element based on voxel data corresponding to the first region, voxel data corresponding to the third region, and the first correction parameter; and obtaining a second brain morphological index related to the second brain element based on voxel data corresponding to the second region, voxel data corresponding to the third region, and the second correction parameter. have.

The brain image analysis apparatus disclosed in the present application includes: an image acquisition unit for acquiring a target brain image; and a controller providing brain image analysis information based on the target brain image, wherein the controller obtains a first correction parameter for correcting a first morphological value related to a first brain element, Obtaining a second correction parameter for correcting a second morphological value related to a second brain element different from a brain element, obtaining a target brain image, and applying the target brain image to a first region corresponding to the first brain element , a third region associated with the first region, the second region, and an inner region of the cranial region by performing segmentation into a plurality of brain regions including a second region corresponding to the second brain element and a cranial region obtaining a first brain morphological index related to the first brain element based on voxel data corresponding to the first region, voxel data corresponding to the third region, and the first correction parameter, and and acquire a second brain morphological index related to the second brain element based on voxel data corresponding to the second region, voxel data corresponding to the third region, and the second correction parameter.

Solutions of the present invention are not limited to the above-described solutions, and solutions not mentioned will be clearly understood by those of ordinary skill in the art to which the present invention belongs from the present specification and the accompanying drawings. will be able

According to the embodiment of the present application, it is possible to increase the reliability of the medical image analysis result by performing quality analysis on the target medical image and providing information on the analysis result to the user.

According to the embodiment of the present application, the morphological index is calculated based on the medical image of the object, and the morphological index is more accurately calculated by appropriately applying the correction parameter in consideration of the scan condition in which the medical image is obtained or the position of the target element. can do.

According to the embodiment of the present application, user convenience can be improved by selectively providing the necessary index information to the user from among various index information obtained through the analysis of the target medical image.

The effects of the present invention are not limited to the above-mentioned effects, and the effects not mentioned will be clearly understood by those of ordinary skill in the art to which the present invention belongs from the present specification and the accompanying drawings.

1 is a schematic diagram of a medical image analysis system according to an embodiment of the present application.
2 is a schematic diagram of an image analysis apparatus 2000, a learning apparatus 2200, a correction parameter obtaining apparatus 2400, and an output apparatus 2600 for image analysis according to an embodiment of the present application
3 is a block diagram of an image analysis apparatus according to an embodiment of the present application.
4 is a block diagram of an output device according to an embodiment of the present application.
5 illustrates an example of an image alignment operation of the image analysis apparatus 2000 .
6 illustrates an example of an image alignment operation of the image analysis apparatus 2000 .
7 is a diagram illustrating a flowchart of a process for image segmentation according to an embodiment of the present application.
8 is a flowchart of a learning method of a neural network model of the learning apparatus 2200 according to an embodiment of the present application.
9 is an exemplary structural diagram of an image data set according to an embodiment of the present application.
10 is an example of an artificial neural network model that can be used by the learning apparatus 2200 according to an embodiment of the present application.
11 is another example of an artificial neural network model that can be used by the learning apparatus 2200 according to an embodiment of the present application.
12 is a flowchart of an image segmentation method using a neural network model of the image analysis apparatus 2000 according to an embodiment of the present application.
13 is an exemplary structural diagram of a target image according to an embodiment of the present application.
14 is an example of images obtained based on a result value obtained by a segmentation process of the image analysis apparatus 2000 according to an embodiment of the present application.
15 is a diagram illustrating a flowchart of a process for image segmentation according to an embodiment of the present application.
16 is an exemplary structural diagram of image data sets according to an embodiment of the present application.
17 is a diagram illustrating a flowchart of a process for image segmentation according to an embodiment of the present application.
18 is a diagram for describing an image quality determination process according to an embodiment.
19 is a diagram for explaining an image analysis apparatus that performs an image quality determination process.
20 is a diagram for explaining an image quality determination module.
21 is a diagram for explaining a first image quality determination process.
22 is a diagram for describing a first image quality determining unit.
23 is a diagram for explaining a second image quality determination process.
24 is a diagram for exemplarily explaining that the second image quality determination process determines the location of the artifact in the medical image using the artificial neural network model.
25 is a diagram for describing a neural network model for acquiring artifact information according to an embodiment.
26 is a diagram for explaining a third image quality determination process.
27 is a diagram for explaining an embodiment of a third image quality determination process.
28 is a diagram for explaining a fourth image quality determination process.
29 is a diagram illustrating an image in which an artifact has occurred.
30 is a diagram for explaining determining whether an artifact is generated to overlap an ROI in a fourth image quality determination process according to another exemplary embodiment;
31 is a diagram for describing an image quality related information output module according to an exemplary embodiment.
32 is a diagram for exemplarily explaining a first error information screen.
33 and 34 are diagrams for exemplarily explaining the second error information screen.
35 and 36 are diagrams for exemplarily explaining a third error information screen.
37 and 38 are diagrams for exemplarily explaining a fourth error information screen.
FIG. 39 is a diagram for exemplarily explaining a selection window output by the image quality related information output module 3900 .
40 is a diagram for illustratively explaining a selection window output by the image quality related information output module 3900 .
41 to 43 are diagrams for explaining first to third quality determination processes.
44 and 45 are diagrams for explaining a fourth quality determination process.
46 and 47 are diagrams for explaining a fifth quality determination process.
48 is a flowchart illustrating an operation of an image analysis method implemented by the image analysis apparatus 2000 according to an embodiment of the present application.
49 is an exemplary diagram of an image alignment method implemented by the image analysis apparatus 2000 according to an embodiment of the present application.
50 is a flowchart of an image analysis method according to an embodiment of the present application.
51 is a diagram illustrating an example of a method of correcting the first boundary of the first inner region of the skull according to an embodiment of the present application.
52 is a diagram illustrating an example of a second internal region obtained according to a method of correcting a first boundary of a first internal region of a skull according to an embodiment of the present application.
53 is a diagram illustrating another example of a method of correcting the first boundary of the first inner region of the skull according to an embodiment of the present application.
54 to 57 are exemplary views illustrating information related to a morphological index output according to an embodiment of the present application.
58 is a flowchart illustrating an operation of a morphological numerical correction method implemented by the image analysis apparatus 2000 according to an embodiment of the present application.
59 is a flowchart of an image analysis method according to an embodiment of the present application.
60 is a flowchart illustrating an operation of a method of obtaining a correction parameter implemented by the apparatus 2400 for obtaining a correction parameter according to an embodiment of the present application.
61 is an exemplary diagram for describing information included in a plurality of image data sets according to an embodiment of the present application.
62 is a graph illustrating an example of correlation analysis for acquiring a correction parameter of the correction parameter acquiring apparatus 2400 according to an embodiment of the present application.
63 is a graph illustrating another example of correlation analysis for obtaining a correction parameter of the apparatus 2400 for obtaining a correction parameter according to an embodiment of the present application.
64 is an exemplary diagram for describing information included in a plurality of image data sets according to an embodiment of the present application.
65 is an exemplary diagram of correlation analysis between morphological values according to a location where a target element is distributed in a skull according to an embodiment of the present application.
66 is a flowchart of an image analysis method according to an embodiment of the present application.
67 is a diagram schematically illustrating a user interface according to an embodiment of the present application.
68 is a diagram for describing a medical information output process according to an embodiment.
69 is a diagram for describing an image analysis apparatus according to an exemplary embodiment.
70 is a diagram for explaining a medical data acquisition module.
71 is a diagram for explaining a medical information acquisition module.
72 is a diagram for describing a diagnostic auxiliary information obtaining module.
73 is a diagram for explaining a diagnostic auxiliary information output module.
74 is a diagram for illustrating related indicators for some target diseases.
75 is a diagram for explaining a medical information output screen according to an exemplary embodiment.
76 is a diagram for explaining a medical information output process according to another embodiment.
77 is a diagram for explaining an image analysis apparatus according to another embodiment.
78 is a diagram for explaining a selection information obtaining module.
79 is a diagram for explaining a selection information output module.
80 and 81 are diagrams for explaining a selection information output screen.

The above-described objects, features and advantages of the present application will become more apparent from the following detailed description in conjunction with the accompanying drawings. However, since the present application may have various changes and may have various embodiments, specific embodiments will be exemplified in the drawings and described in detail below.

Throughout the specification, like reference numerals refer to like elements in principle. In addition, components having the same function within the scope of the same idea shown in the drawings of each embodiment will be described using the same reference numerals, and overlapping descriptions thereof will be omitted.

If it is determined that a detailed description of a known function or configuration related to the present application may unnecessarily obscure the gist of the present application, the detailed description thereof will be omitted. In addition, numbers (eg, first, second, etc.) used in the description process of the present specification are only identification symbols for distinguishing one component from other components.

In addition, the suffixes "module" and "part" for the components used in the following embodiments are given or mixed in consideration of only the ease of writing the specification, and do not have a meaning or role distinct from each other by themselves.

In the following examples, the singular expression includes the plural expression unless the context clearly dictates otherwise.

In the following embodiments, terms such as include or have means that the features or components described in the specification are present, and the possibility of adding one or more other features or components is not excluded in advance.

In the drawings, the size of the components may be exaggerated or reduced for convenience of description. For example, the size and thickness of each component shown in the drawings are arbitrarily indicated for convenience of description, and the present invention is not necessarily limited to the illustrated bar.

In cases where certain embodiments are otherwise implementable, the order of specific processes may be performed differently from the order in which they are described. For example, two processes described in succession may be performed substantially simultaneously, or may be performed in an order opposite to the order described.

In the following embodiments, when components are connected, it includes not only cases in which components are directly connected but also cases in which components are interposed between components and connected indirectly.

For example, in the present specification, when it is said that a component is electrically connected, it includes not only a case in which the component is directly electrically connected, but also a case in which a component is interposed in the middle and electrically connected indirectly.

In a medical image analysis method according to an embodiment of the present application, the method includes: acquiring a target medical image including at least one artifact; obtaining artifact information related to the artifact, based on the target medical image, wherein the artifact information includes an artifact region indicating a region in which the artifact is distributed; Using a first neural network model, based on the target medical image, obtaining a plurality of regions each corresponding to one or more anatomically distinguished elements - The first neural network model includes the plurality of regions based on the medical image is learned to obtain a region of , wherein the plurality of regions include a first region corresponding to the first element, a second region corresponding to the second element, and a third region corresponding to the third element; obtaining a degree of overlap between the artifact region and a region of interest, wherein the region of interest includes at least a portion of the first region; a first quality determination step of determining the quality of the target medical image based on the overlapping degree; and outputting quality-related information of the target medical image based on the first quality determination.

The first element is a skull, and the first region includes a region corresponding to the skull, wherein the region of interest may be the skull region or an inner region of the skull.

The second element may be an element related to diagnosis of a target disease, and the region of interest may include at least a portion of the second region.

The artifact region is obtained using a second neural network model trained to determine whether an artifact is present in the medical image, wherein the artifact region is obtained from the second neural network model and is a target present in the target medical image. Based on the feature map related to the artifact, the relevance to the target artifact on the feature map may be a region equal to or greater than a reference value.

The artifact area includes a first area and a second area positioned within the first area, wherein the first area is an area on the feature map that has a relevance to the target artifact equal to or greater than a first reference value, and the second area may be an area on the feature map that has a relevance to the target artifact equal to or greater than a second reference value greater than the first reference value.

The artifact region may be obtained using a second neural network model trained to segment a region corresponding to the artifact included in the medical image.

The determining of the degree of overlap may include acquiring an outer line of the artifact area and an outer line of the ROI from the target medical image, and the outer line of the artifact area and the outer line of the ROI on the medical image and determining whether the artifact and the region of interest overlap, based on whether they overlap.

The first quality determination step may include determining that the quality of the target medical image is normal when there are less than two intersections between the outer line of the artifact area and the outer line of the ROI overlapping on the target medical image, and the intersection is In the case of two or more, it may include determining that the quality of the target medical image is abnormal.

The determining of the degree of overlap may include: obtaining, on the target medical image, the number of common pixels included in both the artifact region and the region of interest; and determining a degree of overlap between the artifact and the region of interest by using the number of common pixels.

In the first quality determination step, it is determined that the quality of the target medical image is abnormal when the number of common pixels exceeds a predetermined reference value, and when the number of common pixels is less than or equal to a predetermined reference value, the quality of the target medical image is normal It may include judging that

A second quality determination step of determining the quality of the target medical image based on the third area, wherein the second quality determination step includes: obtaining a morphological index based on the third area; The method may include determining whether the morphological index satisfies a quantitative criterion.

The morphological index may be obtained based on the first region and the third region.

The morphological index may be obtained based on a ratio of a volume of the first region to a volume of the third region.

The third region may be a region corresponding to an element positioned at the bottom of the target medical image.

The third region may be a region corresponding to an element related to diagnosis of a target disease.

The outputting the quality-related information of the medical image includes displaying error information generated based on the first quality determination and a selection window related to the error information, wherein the error information includes the artifact information and the plurality of pieces of information. It may contain information about the area of

The selection window includes a first object, and the method for analyzing a medical image includes performing a subsequent operation according to a user's selection of the first object, wherein the subsequent operation includes: the target medical image analysis operation; It may include any one of an operation of correcting a target medical image or an operation of re-photographing the target medical image.

An apparatus for analyzing a medical image according to another embodiment of the present application includes: an acquisition unit configured to acquire a medical image; a processing unit configured to obtain image quality information based on the medical image; and an output unit for outputting the image quality information, wherein the processing unit acquires a target medical image including an artifact through the acquiring unit, and acquires artifact information related to the artifact based on the target medical image and - the artifact information includes an artifact area indicating a region in which the artifact is distributed; obtain a region of - the first neural network model is trained to acquire the plurality of regions based on the medical image, wherein the plurality of regions are a first region corresponding to a first element and a first region corresponding to a second element including a second region and a third region corresponding to a third element, obtaining an overlapping degree of the artifact region and a region of interest, wherein the region of interest includes at least a portion of the first region, the overlapping degree may perform a first quality determination of the target medical image based on , and the output unit may output quality-related information of the target medical image based on the first quality determination.

The first element is a skull, and the first region includes a region corresponding to the skull, wherein the region of interest may be the skull region or an inner region of the skull.

The second element may be an element related to diagnosis of a target disease, and the region of interest may include at least a portion of the second region.

The artifact region is obtained using a second neural network model trained to determine whether an artifact is present in the medical image, wherein the artifact region is obtained from the second neural network model and is a target present in the target medical image. Based on the feature map related to the artifact, the relevance to the target artifact on the feature map may be a region equal to or greater than a reference value.

The artifact area includes a first area and a second area positioned within the first area, wherein the first area is an area on the feature map that has a relevance to the target artifact equal to or greater than a first reference value, and the second area may be an area on the feature map that has a relevance to the target artifact equal to or greater than a second reference value greater than the first reference value.

The artifact region may be obtained using a second neural network model trained to segment a region corresponding to the artifact included in the medical image.

The processing unit may be configured to obtain an outer line of the artifact region and an outer line of the ROI from the target medical image, and based on whether the outer line of the artifact region and the outer line of the ROI overlap on the medical image Thus, it may be determined whether the artifact and the region of interest overlap.

The processing unit is configured to determine that the quality of the target medical image is normal when there are less than two intersections between the outer line of the artifact area and the outer line of the ROI overlapping on the target medical image. It may be determined that the quality of the target medical image is abnormal.

The processing unit may obtain, on the target medical image, the number of common pixels included in both the artifact region and the region of interest, and determine a degree of overlap between the artifact and the region of interest by using the number of common pixels. have.

The processing unit may determine that the quality of the target medical image is abnormal when the number of common pixels exceeds a predetermined reference value, and determine that the quality of the target medical image is normal when the number of common pixels is less than or equal to a predetermined reference value have.

The processing unit may obtain a morphological index based on the third region, and perform a second quality determination based on whether the morphological index satisfies a quantitative criterion.

The morphological index may be obtained based on the first region and the third region.

The morphological index may be obtained based on a ratio of a volume of the first region to a volume of the third region.

The third region may be a region corresponding to an element positioned at the bottom of the target medical image.

The third region may be a region corresponding to an element related to diagnosis of a target disease.

According to the brain image analysis method disclosed in the present application, a brain image including voxel data is obtained; obtaining the reference region and a first boundary defining a first internal region that is inside the cranial region by segmenting the brain image into a plurality of regions including a cranial region and a reference region; obtaining a second inner region having a second boundary by modifying at least a portion of the first boundary associated with the first inner region based on the reference region, a portion of the second inner region being in the first inner region included and comprising the target element; obtaining a first volume value associated with the target element based on voxel data corresponding to the target element; obtaining a second volume value related to the second inner region based on voxel data included in the second boundary; and calculating a volume index of the target element based on the first volume value and the second volume value.

According to the brain image analysis method disclosed in the present application, the method may further include aligning the brain image to obtain the second internal region.

According to the brain image analysis method disclosed in the present application, aligning the brain image may include: acquiring a first feature region and a second feature region from the brain image;

calculating a first feature point from the first feature region and calculating a second feature point from the second feature region; obtaining a photographing direction of the brain image based on the first feature point and the second feature point; and aligning the brain image so that the photographing direction is parallel to a reference direction.

According to the brain image analysis method disclosed in the present application, the first characteristic region is a region corresponding to an anterior commissure, and the second characteristic region is a region corresponding to a posterior commissure, and the first The feature point may be included in a boundary defining a region corresponding to the anterior commissure, and the second feature point may be included in a boundary defining a region corresponding to the posterior commissure.

According to the brain image analysis method disclosed in the present application, aligning the brain image may include: acquiring data related to a direction of the brain image included in the brain image; obtaining a photographing direction of the brain image based on data related to the direction of the brain image; and aligning the brain image so that the photographing direction corresponds to a reference direction.

According to the brain image analysis method disclosed in the present application, acquiring the second internal region may include: acquiring a reference plane adjacent to the reference region from the brain image; and obtaining the second inner area having the second boundary by modifying a part of the first boundary of the first inner area based on the reference plane.

According to the brain image analysis method disclosed in the present application, the reference plane is a plane parallel to a transverse plane of the brain image, and obtaining the second inner region is an upper side of the reference plane. modifying a portion located below the reference plane of the first boundary based on the first inner region located in ; and obtaining the second inner region having the second boundary based on the modified first boundary.

According to the brain image analysis method disclosed in the present application, the reference region may be a region corresponding to the cerebellum, and the reference plane may be adjacent to a lower edge of a boundary defining the corresponding region of the cerebellum.

According to the brain image analysis method disclosed in the present application, the reference region may be a region corresponding to the cerebellum, and the reference plane may be adjacent to a lower edge of a boundary defining the corresponding region of the cerebellum.

According to the brain image analysis method disclosed in the present application, acquiring the second internal region includes: acquiring a first feature region and a second feature region from the brain image; calculating a first feature point from the first feature region and calculating a second feature point from the second feature region; obtaining a reference direction connecting the first feature point and the second feature point based on the first feature point and the second feature point; obtaining a reference plane parallel to the reference direction; and obtaining the second inner area having the second boundary by modifying a part of the first boundary of the first inner area based on the reference plane.

According to the brain image analysis method disclosed in the present application, the reference region may be a region corresponding to the cervical vertebra, and the reference plane may be adjacent to the region corresponding to the cervical vertebra and may be perpendicular to a sagittal plane.

According to the brain image analysis method disclosed in the present application, obtaining the second inner region is based on the first inner region located on the upper side of the reference plane of the reference plane of the first boundary. Corrected the lower part; and obtaining the second inner region having the second boundary based on the modified first boundary.

According to the brain image analysis method disclosed in the present application, the volume index may be defined as a ratio of the first volume value to the second volume value.

According to the brain image analysis method disclosed in the present application, calculating the volume index of the target element obtains a first correction parameter for correcting the first volume value - The first correction parameter determines that the brain image is acquired based on the acquired scan condition or the position of the target element; and calculating the first volume correction value based on the first correction parameter and the first volume value; obtaining a second correction parameter for correcting the second volume value, wherein the second correction parameter is obtained based on a scan condition in which the brain image is obtained or a position of the target element; and calculating the second volume correction value based on the second correction parameter and the second volume value; and calculating the volume index based on the first volume correction value and the second volume correction value.

According to the brain image analysis method disclosed in the present application, the scan conditions under which the brain image is acquired are the magnetic field strength related to the resolution of the brain image of the brain image acquisition device, the manufacturer of the brain image acquisition device, and the brain image acquisition device. It may relate to at least one of the setting parameters related to the shape of the magnetic field to be generated.

According to the brain image analysis apparatus disclosed in the present application, an image acquisition unit for acquiring a brain image; and a controller that provides brain image analysis information based on the brain image, wherein the controller acquires a brain image including voxel data and uses the brain image to a plurality of regions including a skull region and a reference region By performing segmentation into the cranial regions, a first boundary defining a first internal region that is inside the cranial region and the reference region are obtained, and based on the reference region, at least of the first boundary associated with the first internal region. obtaining a second inner region having a second boundary by modifying a part, wherein a part of the second inner region is included in the first inner region and includes the target element; to obtain a first volume value related to the target element based on the first volume value and to obtain a second volume value related to the second inner region based on voxel data included in the second boundary; and calculate a volume index of the target element based on the two volume values.

According to the brain image analysis apparatus disclosed in the present application, the controller may be configured to align the brain image to obtain the second inner region.

According to the brain image analysis apparatus disclosed in the present application, the controller obtains a first feature region and a second feature region from the brain image, calculates a first feature point from the first feature region, and from the second feature region Calculating a second feature point, obtaining a photographing direction of the brain image based on the first characteristic point and the second characteristic point, and aligning the brain image by aligning the brain image so that the photographing direction is parallel to a reference direction can be configured.

According to the brain image analysis apparatus disclosed in the present application, the first characteristic region is a region corresponding to an anterior commissure, and the second characteristic region is a region corresponding to a posterior commissure, and the first The feature point may be included in a boundary defining a region corresponding to the anterior commissure, and the second feature point may be included in a boundary defining a region corresponding to the posterior commissure.

According to the brain image analysis apparatus disclosed in the present application, the controller acquires data related to a direction of the brain image included in the brain image, and a photographing direction of the brain image based on the data related to the direction of the brain image may be configured to obtain and align the brain image by aligning the brain image so that the photographing direction corresponds to a reference direction.

According to the brain image analysis apparatus disclosed in the present application, the controller obtains a reference plane adjacent to the reference region from the brain image, and selects a part of the first boundary of the first internal region based on the reference plane. and modifying to obtain the second inner area having the second boundary, thereby obtaining the second inner area.

According to the brain image analysis apparatus disclosed in the present application, the reference plane is a plane parallel to a transverse plane of the brain image, and the controller is the first modifying a portion of the first boundary located below the reference plane based on the inner area, and obtaining the second inner area having the second boundary based on the modified first boundary, whereby the second may be configured to obtain an inner region.

According to the brain image analysis apparatus disclosed in the present application, the reference region may be a region corresponding to the cerebellum, and the reference plane may be adjacent to a lower edge of a boundary defining the corresponding region of the cerebellum.

According to the brain image analysis apparatus disclosed in the present application, the controller obtains a first feature region and a second feature region from the brain image, calculates a first feature point from the first feature region, and from the second feature region calculating a second feature point, obtaining a reference direction connecting the first feature point and the second feature point based on the first feature point and the second feature point, and obtaining a reference plane parallel to the reference direction, and obtaining the second inner area by modifying a part of the first boundary of the first inner area based on the reference plane to obtain the second inner area having the second boundary.

According to the brain image analysis apparatus disclosed in the present application, the reference region may be a region corresponding to the cervical vertebra, and the reference plane may be adjacent to the region corresponding to the cervical vertebra and may be perpendicular to a sagittal plane.

According to the brain image analysis apparatus disclosed in the present application, the controller, based on the first inner region located on the upper side of the reference plane, the portion located below the reference plane of the first boundary and to obtain the second inner region having the second boundary based on the modified first boundary.

According to the brain image analysis apparatus disclosed in the present application, the controller may be configured to define the volume index as a ratio of the first volume value to the second volume value.

According to the brain image analysis apparatus disclosed in the present application, the controller acquires a first correction parameter for correcting the first volume value, wherein the first correction parameter is a scan condition under which the brain image is obtained or the target element. is obtained based on a position; and calculates the first volume correction value based on the first correction parameter and the first volume value to obtain the first volume correction value, and corrects the second volume value. obtain a second calibration parameter for: the second calibration parameter is obtained based on a scan condition in which the brain image is obtained or the position of the target element; and, according to the second calibration parameter and the second volume value the volume of the target element by calculating the second volume correction value based on the second volume correction value to obtain the second volume correction value, and calculating the volume index based on the first volume correction value and the second volume correction value; It may be configured to calculate an indicator.

According to the brain image analysis apparatus disclosed in the present application, the scan conditions under which the brain image is obtained are the magnetic field strength related to the resolution of the brain image of the brain image acquisition device, the manufacturer of the brain image acquisition device, and the brain image acquisition device. It may relate to at least one of the setting parameters related to the shape of the magnetic field to be generated.

According to the medical image analysis method disclosed in the present application, a first morphological value obtained from a first medical image obtained under a first scan condition and related to a target element, and a second medical image obtained under a second scan condition, and the obtaining a calibration parameter calculated based on a correlation between a second morphological value associated with the target element; acquiring a target medical image acquired under the second scan condition; obtaining a target region related to the target element from the target medical image by segmenting the target medical image into a plurality of regions corresponding to a plurality of elements including the target element; obtaining a target morphological value related to the target element based on voxel data corresponding to the target region; obtaining a corrected morphological value based on the target morphological value and the correction parameter; and outputting a morphological index based on the corrected morphological value.

According to the medical image analysis method disclosed in the present application, the correction parameter includes a parameter for calculating the first morphological value based on the second morphological value, and obtaining the corrected morphological value comprises: and obtaining the corrected morphological value, which is a morphological estimate value of the target element under the first scan condition, based on the target morphological value and the correction parameter.

According to the medical image analysis method disclosed in the present application, the correction parameter includes a parameter related to a linear function for calculating the first morphological value based on the second morphological value, and the corrected morphological value Acquiring the morphological value may include acquiring the corrected morphological value based on the linear function including the target morphological value and the parameter.

According to the medical image analysis method disclosed in the present application, the first scan condition and the second scan condition include a magnetic field strength related to a resolution of a medical image of a medical image acquisition apparatus, a manufacturer of the medical image acquisition apparatus, and the medical image acquisition It may relate to at least one of the setting parameters related to the shape of the magnetic field generated by the device.

According to the medical image analysis method disclosed in the present application, the target medical image is obtained from a first object having a first characteristic, and the correction parameter is the first medical image obtained from a second object having the first characteristic. and obtained from the second medical image, wherein the first characteristic may be related to the subject's age or gender.

According to the medical image analysis method disclosed in the present application, the segmentation may be performed using a neural network provided to acquire a plurality of regions corresponding to a plurality of elements based on the target medical image.

According to the medical image analysis method disclosed in the present application, the method further includes: converting the target medical image acquired under the second scan condition into a medical image corresponding to the image acquired under the first scan condition, wherein the segmentation comprises: This may be performed based on the converted medical image.

According to the medical image analysis method disclosed in the present application, when the first scan condition and the second scan condition are related to a magnetic field strength, the correction parameter is obtained from the second medical image obtained under a second magnetic field strength, and a target a parameter set for transforming the second morphological value associated with an element into the first morphological value associated with a target element obtained from the first medical image obtained under a first magnetic field strength; When obtained under the second magnetic field strength, the corrected morphological value may be obtained based on the target morphological value and the parameter set.

According to an apparatus for analyzing a medical image disclosed in the present application, an image acquisition unit configured to acquire a target medical image; and a controller providing medical image analysis information based on the target medical image. wherein the controller includes: a first morphological value obtained from a first medical image obtained under a first scan condition and related to a target element; Obtaining a correction parameter calculated based on a correlation between related second morphological values, obtaining a target medical image obtained under the second scan condition, and generating the target medical image including the target element. By performing segmentation into a plurality of regions corresponding to a plurality of elements, a target region related to the target element is obtained from the target medical image, and the target element is based on voxel data corresponding to the target region. obtain a target morphological value associated with .

According to the medical image analysis apparatus disclosed in the present application, the correction parameter includes a parameter for calculating the first morphological value based on the second morphological value, and the controller includes: the target morphological value and the The corrected morphological value, which is a morphological estimate value of the target element under the first scan condition, may be obtained based on the correction parameter.

According to the medical image analysis apparatus disclosed in the present application, the correction parameter includes a parameter related to a linear function for calculating the first morphological value based on the second morphological value, and the controller includes: and obtain the corrected morphological value based on the linear function including the morphological value and the parameter.

According to the medical image analysis apparatus disclosed in the present application, the first scan condition and the second scan condition may include a magnetic field strength related to a resolution of a medical image of the medical image acquisition device, a manufacturer of the medical image acquisition device, and the medical image acquisition device. It may relate to at least one of a setting parameter setting parameter related to the shape of the magnetic field generated by the device.

According to the medical image analysis apparatus disclosed in the present application, the target medical image is obtained from a first object having a first characteristic, and the correction parameter is the first medical image obtained from a second object having the first characteristic and obtained from the second medical image, wherein the first characteristic may be related to the subject's age or gender.

According to the medical image analysis apparatus disclosed in the present application, the segmentation may be performed using a neural network provided to acquire a plurality of regions corresponding to a plurality of elements based on the target medical image.

According to the medical image analysis apparatus disclosed in the present application, the controller converts the target medical image acquired under the second scan condition into a medical image corresponding to the image acquired under the first scan condition, and the segmentation is It may be configured to be performed based on the converted medical image.

According to the medical image analysis apparatus disclosed in the present application, the apparatus further includes an output module for outputting the morphological index and morphological information obtained based on the morphological index database, wherein the morphological information includes the morphological index database and the morphological index database. Percentile information on the morphological index database of the morphological index related to the object of the target medical image is included based on the morphological index, and the output module outputs the morphological information reflecting the percentile information. can be configured.

According to the medical image analysis apparatus disclosed in the present application, when the first scan condition and the second scan condition are related to the magnetic field strength, the correction parameter is obtained from the second medical image obtained under the second magnetic field strength, and the target a parameter set for transforming the second morphological value associated with an element into the first morphological value associated with a target element obtained from the first medical image obtained under a first magnetic field strength; When obtained under the second magnetic field strength, the corrected morphological value may be obtained based on the target morphological value and the parameter set.

According to the medical image analysis method disclosed in the present application, a target medical image is obtained; obtaining a target scan condition related to the target medical image; obtaining a target morphological value related to the target element based on voxel data of a target region corresponding to the target element included in the target medical image; determining a target calibration parameter among one or more calibration parameters based on the target scan condition; and outputting a morphological index based on the determined target correction parameter and the target morphological value, wherein the determining of the target correction parameter includes, when the target scan condition corresponds to the first scan condition, The target correction parameter is determined as a first correction parameter for correcting a first morphological value related to a target element obtained under a first scan condition, and when the target scan condition corresponds to the second scan condition, the target The correction parameter may include determining as a second correction parameter for correcting a second morphological value related to the target element obtained under a second scan condition different from the first scan condition.

According to the medical image analysis method disclosed in the present application, outputting the morphological index may include, when the target scan condition corresponds to the first scan condition, based on the target morphological value and the first correction parameter. 1 calculate a corrected morphological index, wherein when the target scan condition corresponds to the second scan condition, the first corrected morphological index is different from the first corrected morphological index based on the target morphological value and the second calibration parameter. calculating a second corrected morphological index; and outputting the first corrected morphological index when the target scan condition corresponds to the first scan condition, and outputting the second corrected morphological index when the target scan condition corresponds to the second scan condition. Output box; may include.

According to the medical image analysis method disclosed in the present application, a reference scan condition as a criterion for determining the target correction parameter is obtained; may further include.

According to the medical image analysis method disclosed in the present application, the first correction parameter includes a parameter for calculating a morphological value obtained under the first scan condition as a morphological estimate value under a third scan condition, and the second correction parameter The parameter includes a parameter for calculating a morphological value obtained under the second scan condition as a morphological estimate value under a third scan condition, wherein calculating the first corrected morphological index comprises: the target morphological value and obtaining, based on the first calibration parameter, the first corrected morphological value that is a morphological estimate of the target element under the second scan condition, wherein calculating the second corrected morphological index comprises: , obtaining the second corrected morphological value, which is a morphological estimate value of the target element under the third scan condition, based on the target morphological value and the second correction parameter.

According to the medical image analysis method disclosed in the present application, the first correction parameter is a first parameter related to a first linear function for calculating the morphological value obtained under the first scan condition as a morphological estimate value under the third scan condition. a set, wherein the second calibration parameter comprises a second set of parameters associated with a second linear function for calculating a morphological value obtained under the second scan condition as a morphological estimate under a third scan condition, wherein calculating a first corrected morphological index comprises obtaining the first corrected morphological value based on the first linear function including the target morphological value and the first parameter set; Calculating the second corrected morphological index may include obtaining the second corrected morphological value based on the second linear function including the target morphological value and the second parameter set. can

According to the medical image analysis method disclosed in the present application, the first scan condition, the second scan condition, and the target scan condition include a magnetic field strength related to a resolution of a medical image of a medical image acquisition device, and a manufacturer of the medical image acquisition device and a setting parameter related to a shape of a magnetic field generated by the medical image acquisition device.

According to the medical image analysis method disclosed in the present application, the target medical image is obtained from an object having a first characteristic, and the first correction parameter or the second correction parameter is obtained from the object having the first characteristic. It is obtained from the first medical image and the second medical image, and the first characteristic may be related to the age or gender of the object.

According to the medical image analysis method disclosed in the present application, obtaining the target morphological value may include segmenting the target medical image into a plurality of regions corresponding to a plurality of elements including at least the target element, obtaining a target area related to the target element from the target medical image; and obtaining the target morphological value based on voxel data corresponding to the target area, wherein the segmentation is a neural network provided to obtain a plurality of areas corresponding to a plurality of elements based on the target medical image. can be performed using

According to the medical image analysis method disclosed in the present application, the method further includes: converting the target medical image acquired under the target scan condition into a medical image corresponding to the image captured under a scan condition other than the target scan condition; Segmentation may be performed based on the converted medical image.

According to the medical image analysis method disclosed in the present application, when the first scan condition, the second scan condition, and the target scan condition are related to the magnetic field strength, the first correction parameter is the target obtained under the first magnetic field strength. a first parameter set for correcting the first morphological value associated with an element, wherein the second correction parameter comprises a first parameter set for correcting the second morphological value associated with the target element obtained under a second magnetic field strength. two parameter sets, wherein the determining of the target calibration parameter includes: when the target scan condition corresponds to the first magnetic field strength, the target calibration parameter is determined as the first parameter set, and the target scan condition is When it corresponds to the second magnetic field strength, the target correction parameter may include determining a second parameter set.

According to an apparatus for analyzing a medical image disclosed in the present application, an image acquisition unit configured to acquire a target medical image; and a controller providing medical image analysis information based on the target medical image. wherein the controller obtains a target scan condition related to the target medical image, and a target morphological value related to the target element based on voxel data of a target region corresponding to the target element included in the target medical image. and determine a target correction parameter among one or more correction parameters based on the target scan condition, and output a morphological index based on the determined target correction parameter and the target morphological value, wherein the target scan condition is configured to: When the first scan condition corresponds to the first scan condition, the target calibration parameter is determined as a first calibration parameter for correcting a first morphological value related to a target element obtained under the first scan condition, wherein the target scan condition is By determining the target correction parameter as a second correction parameter for correcting a second morphological value associated with the target element obtained under a second scan condition that is different from the first scan condition, if it corresponds to the second scan condition and determine the target calibration parameter.

According to the medical image analysis apparatus disclosed in the present application, when the target scan condition corresponds to the first scan condition, the controller is configured to configure the first corrected morphology based on the target morphological value and the first correction parameter. calculate an index, and when the target scan condition corresponds to the second scan condition, the second corrected shape different from the first corrected morphological index based on the target morphological value and the second correction parameter calculate a morphological index, output the first corrected morphological index when the target scan condition corresponds to the first scan condition, and output the second corrected morphological index when the target scan condition corresponds to the second scan condition It may be configured to output the morphological index by outputting the morphological index.

According to the medical image analysis apparatus disclosed in the present application, the first correction parameter includes a parameter for calculating a morphological value obtained under the first scan condition as a morphological estimate value under a third scan condition, and the second correction parameter The parameters include parameters for calculating the morphological value obtained under the second scan condition as a morphological estimate value under the third scan condition, wherein the controller is configured to: based on the target morphological value and the first correction parameter, calculating the first corrected morphological index by obtaining the first corrected morphological value, which is a morphological estimate value of the target element under the second scan condition, based on the target morphological value and the second correction parameter to calculate the second corrected morphological index by obtaining the second corrected morphological value that is a morphological estimate value of the target element under the third scan condition.

According to the medical image analysis apparatus disclosed in the present application, the first correction parameter is a first parameter related to a first linear function for calculating the morphological value obtained under the first scan condition as a morphological estimate value under the third scan condition. a set, wherein the second calibration parameter comprises a second set of parameters associated with a second linear function for calculating a morphological value obtained under the second scan condition as a morphological estimate under a third scan condition, wherein a controller, based on the first linear function including the target morphological value and the first parameter set, obtains the first corrected morphological value to calculate the first corrected morphological index; and calculate the second corrected morphological index by obtaining the second corrected morphological value based on the second linear function including a target morphological value and the second parameter set.

According to the medical image analysis apparatus disclosed in the present application, the first scan condition and the second scan condition may include a magnetic field strength related to a resolution of a medical image of the medical image acquisition device, a manufacturer of the medical image acquisition device, and the medical image acquisition device. It may relate to at least one of a setting parameter setting parameter related to the shape of the magnetic field generated by the device.

According to the medical image analysis apparatus disclosed in the present application, the target medical image is obtained from an object having a first characteristic, and the target correction parameter including at least one of the first correction parameter and the second correction parameter may include: It is obtained from the first medical image and the second medical image obtained from the object having a first characteristic, wherein the first characteristic may be related to an age or a gender of the object.

According to the medical image analysis apparatus disclosed in the present application, the controller performs segmentation of the target medical image into a plurality of regions corresponding to a plurality of elements including at least the target element, thereby extracting the target medical image from the target medical image. acquire a target region related to a target element, and acquire the target morphological value based on voxel data corresponding to the target region, wherein the segmentation includes a plurality of elements corresponding to a plurality of elements based on the target medical image. can be performed using a neural network provided to obtain regions of

According to the medical image analysis apparatus disclosed in the present application, the controller converts the target medical image acquired under the target scan condition into a medical image corresponding to an image captured under a scan condition other than the target scan condition, and the conversion It may be configured to perform the segmentation based on the obtained medical image.

According to the medical image analysis apparatus disclosed in the present application, when the first scan condition, the second scan condition, and the target scan condition are related to the magnetic field strength, the first correction parameter is the target obtained under the first magnetic field strength. a first parameter set for correcting the first morphological value associated with an element, wherein the second correction parameter comprises a first parameter set for correcting the second morphological value associated with the target element obtained under a second magnetic field strength. two parameter sets, wherein when the target scan condition corresponds to the first magnetic field strength, the target correction parameter is determined as the first parameter set, and the target scan condition is the second magnetic field strength , the target correction parameter may be configured to determine the target correction parameter by determining the second parameter set.

According to the brain image analysis method disclosed in the present application, a first correction parameter is obtained for correcting a first morphological value related to a first brain element; obtaining a second correction parameter for correcting a second morphological value associated with a second brain element different from the first brain element; acquiring a target brain image; The first region by segmenting the target brain image into a plurality of brain regions including a first region corresponding to the first brain element, a second region corresponding to the second brain element, and a skull region; obtaining a third region associated with the second region and an inner region of the cranial region; acquiring a first brain morphological index related to the first brain element based on voxel data corresponding to the first region, voxel data corresponding to the third region, and the first correction parameter; and obtaining a second brain morphological index related to the second brain element based on voxel data corresponding to the second region, voxel data corresponding to the third region, and the second correction parameter. have.

According to the brain image analysis method disclosed in the present application, the method further includes: acquiring a reference morphological value related to the inner region based on voxel data corresponding to the third region, and acquiring the first brain morphological index The method may include: acquiring a first target morphological value associated with the first brain element based on voxel data corresponding to the first region; calculating a first corrected morphological value associated with a first brain element based on the first target morphological value and the first correction parameter; and calculating the first brain morphological index based on the first corrected morphological value and the reference morphological value, wherein obtaining the second brain morphological index comprises: obtaining a second target morphological value associated with the second brain element based on the corresponding voxel data; calculating a second corrected morphological value associated with a second brain element based on the second target morphological value and the second correction parameter; and calculating the second brain morphological index based on the second corrected morphological value and the reference morphological value.

According to the brain image analysis method disclosed in the present application, the first morphological value and the second morphological value are values obtained under a first scan condition, and the target brain image is obtained under the first scan condition, the first calibration parameter comprises a first parameter for calculating a third morphological value associated with the first brain element obtained under a second scan condition different from the first scan condition based on the first morphological value and the second correction parameter includes a second parameter for calculating a fourth morphological value related to the second brain element obtained under the second scan condition based on the second morphological value, Calculating a 1 corrected morphological value comprises: based on the first target morphological value and the first correction parameter, the first corrected morphology that is a morphological estimate of the first brain element under the second scan condition. obtaining a morphological value, wherein calculating the second corrected morphological value comprises: based on the second target morphological value and the second correction parameter, the second brain element under the second scan condition. and obtaining the second corrected morphological value that is a morphological estimate of .

According to the brain image analysis method disclosed in the present application, the first correction parameter includes a first parameter set related to a first linear function for calculating the third morphological value based on the first morphological value, , wherein the second correction parameter comprises a second parameter set associated with a second linear function for calculating the fourth morphological value based on the second morphological value, wherein the first corrected morphological value is Calculating comprises obtaining the first corrected morphological value based on the first target morphological value and a first parameter set associated with the first linear function, wherein the second corrected morphological value is Calculating may include obtaining the second corrected morphological value based on the second target morphological value and a second parameter set associated with the second linear function.

According to the brain image analysis method disclosed in the present application, the first region corresponding to the first brain element may be located adjacent to the skull region compared to the second region corresponding to the second brain element .

According to the brain image analysis method disclosed in the present application, the first brain element is related to an element performing a first brain function, and the second brain element performs a second brain function different from the first brain function. It can be related to elements.

According to the brain image analysis method disclosed in the present application, the first scan condition and the second scan condition are the magnetic field strength related to the resolution of the brain image of the brain image acquisition device, the manufacturer of the brain image acquisition device, and the brain image acquisition It may relate to at least one of the setting parameters related to the shape of the magnetic field generated by the device.

According to the brain image analysis method disclosed in the present application, the target brain image is obtained from a first object having a first characteristic, and the first correction parameter and the second correction parameter are the second object having the first characteristic. It is obtained from a first brain image and a second brain image obtained from

According to the brain image analysis method disclosed in the present application, the segmentation is a plurality of brains corresponding to a plurality of brain elements including the first brain element, the second brain element, and the skull based on the target brain image. This may be done using a neural network provided to acquire regions.

According to the brain image analysis method disclosed in the present application, pre-processing the target brain image; further comprising, wherein the pre-processing of the target brain image includes performing pre-processing on the first region by a first method, It includes performing preprocessing of a second method different from the first method with respect to the two regions, and the segmentation may be performed based on the preprocessed target brain image.

According to the brain image analysis method disclosed in the present application, the first corrected morphological value is a volume value related to the first brain element obtained by the target brain image, and the second target morphological value is the target brain image is a volume value associated with the second brain element obtained as It may be calculated as the second corrected morphological value with respect to the reference morphological value.

According to the brain image analysis apparatus disclosed in the present application, an image acquisition unit for acquiring a target brain image; and a controller providing brain image analysis information based on the target brain image, wherein the controller obtains a first correction parameter for correcting a first morphological value related to a first brain element, Obtaining a second correction parameter for correcting a second morphological value related to a second brain element different from a brain element, obtaining a target brain image, and applying the target brain image to a first region corresponding to the first brain element , a third region associated with the first region, the second region, and an inner region of the cranial region by performing segmentation into a plurality of brain regions including a second region corresponding to the second brain element and a cranial region obtaining a first brain morphological index related to the first brain element based on voxel data corresponding to the first region, voxel data corresponding to the third region, and the first correction parameter, and and acquire a second brain morphological index related to the second brain element based on voxel data corresponding to the second region, voxel data corresponding to the third region, and the second correction parameter.

According to the brain image analysis apparatus disclosed in the present application, the controller is configured to acquire a reference morphological value related to the inner region based on voxel data corresponding to the third region, and a voxel corresponding to the first region obtain a first target morphological value associated with the first brain element based on data, and a first corrected morphological value associated with the first brain element based on the first target morphological value and the first correction parameter and calculating the first brain morphological index based on the first corrected morphological value and the reference morphological value to obtain the first brain morphological index, and voxels corresponding to the second region obtain a second target morphological value associated with the second brain element based on data, and a second corrected morphological value associated with a second brain element based on the second target morphological value and the second correction parameter and calculating the second brain morphological index based on the second corrected morphological value and the reference morphological value to obtain the second brain morphological index.

According to the brain image analysis apparatus disclosed in the present application, the first morphological value and the second morphological value are values obtained under a first scan condition, the target brain image is obtained under the first scan condition, and the the first calibration parameter comprises a first parameter for calculating a third morphological value associated with the first brain element obtained under a second scan condition different from the first scan condition based on the first morphological value and the second correction parameter includes a second parameter for calculating a fourth morphological value related to the second brain element obtained under the second scan condition based on the second morphological value, wherein the controller calculates the first corrected morphological value, which is a morphological estimate value of the first brain element under the second scan condition, based on the first target morphological value and the first correction parameter, and the second target and calculate the second corrected morphological value, which is a morphological estimate of the second brain element under the second scan condition, based on the morphological value and the second correction parameter.

According to the brain image analysis apparatus disclosed in the present application, the first correction parameter includes a first parameter set related to a first linear function for calculating the third morphological value based on the first morphological value, , wherein the second correction parameter comprises a second parameter set associated with a second linear function for calculating the fourth morphological value based on the second morphological value, wherein the first corrected morphological value is Calculating comprises obtaining the first corrected morphological value based on the first target morphological value and a first parameter set associated with the first linear function, wherein the controller is configured to: calculate the first corrected morphological value based on a first set of parameters associated with the first linear function, and the second target morphological value based on the second target morphological value and a second set of parameters associated with the second linear function. to calculate the second corrected morphological value.

According to the brain image analysis apparatus disclosed in the present application, the first region corresponding to the first brain element may be located adjacent to the skull region compared to the second region corresponding to the second brain element .

According to the brain image analysis apparatus disclosed in the present application, the first brain element is related to an element performing a first brain function, and the second brain element performs a second brain function different from the first brain function. It can be related to elements.

According to the brain image analysis apparatus disclosed in the present application, the first scan condition and the second scan condition are the magnetic field strength related to the resolution of the brain image of the brain image acquisition device, the manufacturer of the brain image acquisition device, and the brain image acquisition It may relate to at least one of a setting parameter setting parameter related to the shape of the magnetic field generated by the device.

According to the brain image analysis apparatus disclosed in the present application, the controller, the target brain image is obtained from a first object having a first characteristic, the first correction parameter and the second correction parameter, the first characteristic Branches are obtained from the first brain image and the second brain image obtained from the second object, wherein the first characteristic may be related to the age or sex of the object.

According to the brain image analysis apparatus disclosed in the present application, the segmentation is a plurality of brains corresponding to a plurality of brain elements including the first brain element, the second brain element, and the skull based on the target brain image. This may be done using a neural network provided to acquire regions.

According to the brain image analysis apparatus disclosed in the present application, the controller pre-processes the target brain image, and the segmentation is performed based on the pre-processed target brain image, but the first region is pre-processed by a first method However, the second region may be configured to perform a different pre-processing from the first method.

According to the brain image analysis apparatus disclosed in the present application, the first corrected morphological value is a volume value related to the first brain element obtained by the target brain image, and the second target morphological value is the target brain image is a volume value associated with the second brain element obtained as It may be calculated as the second corrected morphological value with respect to the reference morphological value.

Hereinafter, a brain image analysis apparatus, a brain image analysis system, and a brain image analysis method of the present application will be described. Here, the brain image analysis may be performed by acquiring a plurality of brain regions by performing segmentation on the brain image, calculating a morphological index related to the brain, and outputting it to the user.

In the present application, in order to improve the accuracy and reliability of such brain image analysis, the reliability of the brain image analysis result is improved by checking the quality of the brain image, or the morphological value or morphological index of a specific brain region obtained is corrected. User-friendly report by improving the accuracy of brain image analysis results, calculating morphological values or morphological indicators of specific brain regions from the brain image to be analyzed, providing personalized analysis results, or providing users with only selective information Various techniques are disclosed, such as providing

Hereinafter, a medical image analysis method, a medical image analysis apparatus, and a medical image analysis system according to an embodiment of the present application will be described.

1 is a schematic diagram of a medical image analysis system according to an embodiment of the present application. Referring to FIG. 1 , a medical image analysis system according to an embodiment of the present application may include an image acquisition apparatus 1000 and an image analysis apparatus 2000 .

The image acquisition apparatus 1000 may acquire an image and transmit it to the image analysis apparatus 2000 through a network.

As an example, the image acquisition apparatus 1000 may be an apparatus for acquiring magnetic resonance imaging (MRI). In this case, the magnetic resonance image acquired by the image acquisition apparatus 1000 may be transmitted to the image analysis apparatus 2000 through a network.

As another example, the image acquisition apparatus 1000 may be an apparatus for acquiring a computed tomography (CT) image. In this case, the computed tomography image acquired by the image acquisition apparatus 1000 may be transmitted to the image analysis apparatus 2000 through a network.

As another example, the image obtaining apparatus 1000 may be an apparatus for obtaining an image obtained by radiography. In this case, the radiographic image acquired by the image acquisition apparatus 1000 may be transmitted to the image analysis apparatus 2000 through a network.

As another example, the image acquisition apparatus 1000 may be an apparatus for acquiring an image acquired by positron emission tomography (PET).

However, the above-described image acquisition device 1000 is merely an example, and is not limited thereto, and should be interpreted as encompassing any suitable device or system used for medical imaging.

The image acquired by the image acquisition apparatus 1000 may be a two-dimensional image. In this case, the image may include pixel information related to pixel coordinates, color, intensity, and the like.

The image acquired by the image acquisition apparatus 1000 may be a three-dimensional image. In this case, the image may include pixel information related to the coordinates, color, and intensity of the voxel.

The image acquired by the image acquisition apparatus 1000 may include information related to image alignment. For example, the image acquisition apparatus 1000 may also acquire data ijk related to the direction of the photographed image in consideration of the direction RAS of the reference coordinate axis of the object 100 . Specifically, the image acquisition device acquires the data ijk related to the direction of the photographed image in consideration of the information (xyz) on the coordinate axis of the image acquisition device and the information (RAS) on the reference coordinate axis of the object 100 . can

The image acquisition apparatus 1000 may acquire data related to the magnetic field strength of the image, data related to the manufacturer of the image device, data related to setting parameters of the image device, data related to the object of the image, and the like.

The data related to the magnetic field strength may include data related to the strength of the magnetic field applied to the object when a medical image is acquired using a magnetic field (eg, MRI). For example, when the target image is acquired under a condition of a magnetic field strength of 1.5T (Tesla), data corresponding to “1.5T” may correspond to data related to the magnetic field strength. However, it is not limited to 1.5T, and data corresponding to the magnetic field strength generally used to obtain a medical image, such as 3T, 7T, 8T, etc., may correspond to the data related to the magnetic field strength.

The setting parameters of the imaging device may include parameters that can be adjusted or controlled by the imaging device in order to acquire a medical image. For example, the setting parameters of the imaging device include a repetition time (TR, Repetition Time), which means the time between consecutive pulse sequences applied to the same slice, a time between transmitting an RF pulse and receiving an echo signal (TE, Time to Echo), A time constant related to the rate at which excited protons return to equilibrium (T1), a time constant related to the rate at which excited protons reach equilibrium or out of phase with each other (T2), proton density (PD, Proton density), or any of these It may mean encompassing a combination of

However, it is not limited to the above-described examples and should be interpreted as meaning including any parameters related to magnetic field characteristics. In addition, it is not limited to acquiring a target image using a magnetic field, and should be interpreted as meaning including any parameters related to an imaging device using a CT method or a radiation (eg, X-ray) irradiation method.

In this case, the above-described data may be processed as metadata in the acquired image and transmitted to the image analysis apparatus 2000 or may be transmitted to the image analysis apparatus 2000 separately from the image.

The image acquisition apparatus 1000 may be an apparatus for acquiring medical images such as MRI, CT, and X-ray.

The image acquisition apparatus 1000 may acquire images under various scan conditions.

For example, when the imaging device is an MRI, an image may be acquired under a scanning condition in which the magnetic field strength is 1.5 Tesla (hereinafter, T). Also, an image may be acquired under a scan condition in which the magnetic field strength is 3T. Also, an image may be acquired under a scan condition in which the magnetic field strength is 7T or 8T.

For another example, when the image acquisition apparatus 1000 is an MRI, as described above, the repetition time (TR), time to echo (TE), proton density (PD), or excited protons of the image acquisition apparatus 1000 are Under the scan conditions of a set parameter consisting of a time constant (T1) related to the rate of return to equilibrium, a time constant (T2) related to the rate at which excited protons reach equilibrium or out of phase with each other, or any combination thereof can be obtained.

As another example, the image acquisition apparatus 1000 may be a CT-related apparatus, and in this case, a voltage, a current, an exposure time, a scanning time ( An image may be acquired under scanning conditions of setting parameters related to scanning time), projection, and any combinations thereof.

As another example, the image acquisition apparatus 1000 may be an X-ray-related apparatus, and in this case, the tube voltage, tube current, imaging time, and distance between the X-ray tube apparatus and the detector that can be set in the X-ray image apparatus. Images can be acquired under scan conditions of setting parameters related to (SID), the angle of the x-ray tube supporter, and any combination thereof.

In addition, the image may be obtained differently depending on the tendency of the operator (applicator) to irradiate the radiation. In this case, information on the propensity of the operator may be parameterized. In addition, the parameter for the operator's propensity may be used as a factor to consider in relation to correcting a morphological value or a morphological index in consideration of a scan condition to be described later. In other words, the morphological value or morphological index may be corrected according to the propensity of the operator. In more detail, the image analysis apparatus 2000 according to an embodiment may acquire the operator's identification information in relation to the scan condition. Also, the image analysis apparatus 2000 may be implemented to acquire a correction parameter for correcting a morphological value or a morphological index based on the operator's identification information. For example, the image analysis apparatus 2000 may determine the correlation between the first morphological value calculated from the first image photographed by the first operator and the second morphological value calculated from the second image photographed by the second operator. Based on the correction parameter for estimating the morphological value calculated from the target image obtained from the first operator as the morphological value associated with the second operator (or the morphological value calculated from the target image obtained from the second operator) 1) a correction parameter for estimating the morphological value associated with the operator) may be obtained.

As another example, the image acquisition apparatus 1000 may be an MRI apparatus or a CT apparatus using positron emission tomography (PET). When PET imaging is used, drugs related to radiopharmaceuticals may be used. For example, a drug (or tracer) such as 18F-florbetapir, 18F-florbetaben, 18F-flutemetamol, 18F-Florapronol, etc. may be used to measure amyloid beta. In order to measure the tau protein, a drug (or tracer) such as 18F-flortaucipir may be used. The drug used may be used differently depending on the manufacturer of the image acquisition device. At this time, information on the drug used may be parameterized. A parameter for a drug used for the image acquisition device may be used as a scan condition, and may be used as a factor to consider in relation to correcting a morphological value or a morphological index in consideration of a scan condition to be described later. In other words, the morphological value or the morphological index may be corrected according to the manufacturer of the image acquisition device and the drug used. More specifically, the image analysis device 2000 may acquire information on the manufacturer of the image acquisition device and the drug used, and calculates a morphological value or morphological index based on the information on the manufacturer of the image acquisition device and the drug used. It may be implemented to obtain a correction parameter for correcting. For example, the image analysis apparatus 2000 uses a first morphological value calculated from a first image taken using a first drug (eg, 18F-florbetapir) and a second drug (eg, 18F-florbetaben). Estimating the morphological value calculated from the target image obtained using the first drug as the morphological value obtained using the second drug based on the correlation between the second morphological values calculated from the second image A first correction parameter for (or a correction parameter for estimating a morphological value calculated from a target image obtained by using the second drug as a morphological value obtained by using the first drug) may be obtained. In addition, the image analysis apparatus 2000 provides a correlation between the second morphological value calculated from the second image photographed using the second drug and the third morphological value calculated from the third image photographed using the third drug Based on the relationship, a second calibration parameter for estimating the morphological value calculated from the target image obtained using the second drug as the morphological value obtained using the third drug (eg, 18F-flutemetamol) (or A correction parameter for estimating the morphological value calculated from the target image obtained using the third drug as the morphological value obtained using the second drug) may be obtained. In this case, the first correction parameter and the second correction parameter may be different from each other.

Also, the image acquisition apparatus 1000 may acquire an image under a scan condition including at least one or more setting parameters. In other words, an image may be acquired under scan conditions based on various combinations of the above-described setting parameters.

For example, an image may be acquired under a first scan condition in which the setting parameters are a first combination. Also, an image may be acquired under a second scan condition in which the setting parameter is the second combination and the magnetic field strength is the first strength. The present invention is not limited thereto, and images acquired by the image acquiring apparatus 1000 may be images acquired under various scanning conditions according to various combinations of setting parameters.

Also, the image acquisition apparatus 1000 may be configured with a plurality of image acquisition apparatuses 1000 .

In this case, the plurality of image acquisition devices 1000 may be devices manufactured by different manufacturers. Images acquired by devices manufactured by different manufacturers may have different characteristics, such as brightness and intensity, even if scanning conditions and setting parameters are the same. Therefore, even when a medical image is obtained for the same object, morphological indices based on the medical image may be different depending on the manufacturer's device.

Accordingly, a correction parameter acquisition operation for controlling a scan condition, a setting parameter, or a variable according to a manufacturer of an image device is required by the correction parameter acquiring device 2400 according to an embodiment of the present application, which will be described later.

The image acquired by the image acquisition apparatus 1000 may include information related to an anatomical structure of a specific part of the body. In addition, a specific part of the body may correspond to all parts where medical imaging can be utilized. In the drawings and specifications to be described later for convenience of explanation, the images related to the brain are mainly described, but this is only an example and the various embodiments disclosed in the present application are any suitable body parts other than the brain (for example, , lung, breast, heart, joint, blood vessel, etc.).

Meanwhile, the image acquisition apparatus 1000 according to an embodiment of the present application may be implemented in the form of a server. In this case, the server may be configured to store the medical image and information related to the medical image. In addition, the server may be implemented to modify or process a medical image and information related to the medical image.

Also, the medical images may be stored in a memory or a server of the image analysis apparatus 2000 and used to calculate a correction parameter, perform quality control, or output an analysis result. In this regard, it will be described later in detail below.

Hereinafter, an image analysis apparatus 2000 , a learning apparatus 2200 , a correction parameter obtaining apparatus 2400 , and an output apparatus 2600 for image analysis according to an embodiment of the present application will be described with reference to FIG. 2 . 2 is a schematic diagram of an image analysis apparatus 2000, a learning apparatus 2200, a correction parameter obtaining apparatus 2400, and an output apparatus 2600 for analyzing a medical image according to an embodiment of the present application.

The image analysis apparatus 2000 performs an operation of segmenting the image obtained from the image obtaining apparatus 1000 using an artificial neural network learned by the learning apparatus 2200 or calculating a morphological index of a target element included in the image. can be done

The learning apparatus 2200 may perform an operation of learning a neural network model for image segmentation or a neural network model for image quality control, using a plurality of image data sets.

The correction parameter obtaining apparatus 2400 uses a plurality of image data sets and data related to the scan condition to obtain a morphological value or morphological value related to a target element of the target image (or target image) obtained by the image analysis apparatus 2000 . An operation of calculating a correction parameter for correcting the indicator may be performed.

The image analysis apparatus 2000, the learning apparatus 2200, and the correction parameter obtaining apparatus 2400 shown in FIG. 2 may be implemented to transmit and receive data to each other using any communication method.

For example, the image analysis apparatus 2000 , the learning apparatus 2200 , and the correction parameter obtaining apparatus 2400 may be implemented to share a server.

As shown in FIG. 2 , the image analysis apparatus 2000 , the learning apparatus 2200 , and the correction parameter acquisition apparatus 2400 are illustrated as being provided as separate apparatuses, but this is only an example, and the image analysis apparatus 2000 is provided as separate apparatuses. ), the learning device 2200 and/or the correction parameter obtaining device 2400 may be implemented as a single device. Alternatively, some of the image analysis device 2000 , the learning device 2200 , and the correction parameter acquisition device 2400 may be provided as separate devices, and the remaining devices may be implemented as a single device.

Hereinafter, the configuration of the image analysis apparatus 2000 according to an embodiment of the present application will be described with reference to FIG. 3 . 3 is a block diagram of an image analysis apparatus 2000 according to an embodiment of the present application.

The image analysis apparatus 2000 according to an embodiment of the present application may include a first communication module 2010 , a first memory 2020 , and a first controller 2030 .

The first communication module 2010 may communicate with any external device including the image acquisition device 1000 , the learning device 2200 , the correction parameter acquisition device 2400 , and an output device 2600 to be described later. In other words, the image analysis device 2000 transmits/receives an image from the image acquisition device 1000 through the first communication module 2010 or the learning device 2200 , the parameter acquisition device 2400 , and an output device 2600 to be described later. ), it can transmit and receive data with external devices including repeaters and servers.

For example, the image analysis apparatus 2000 obtains an image obtained from the image obtaining apparatus 1000 through the first communication module 2010 , information about a neural network model learned from the learning apparatus 2200 , and a correction parameter Information related to the calculated calibration parameter may be received from the device 2400 . As another example, the image analysis apparatus 2000 transmits, through the first communication module 2010, information related to a scan condition of an image to the correction parameter obtaining apparatus 2400 or outputs information related to an analysis result. (2600). As another example, the image analysis apparatus 2000 may access the Internet through the first communication module 2010 and upload various data related to images, information related to scan conditions, and information related to analysis results.

The first communication module 2010 is largely divided into a wired type and a wireless type. Since the wired type and the wireless type have their respective advantages and disadvantages, in some cases, the image analysis apparatus 2000 may be provided with both the wired type and the wireless type at the same time.

Here, in the case of the wired type, LAN (Local Area Network) or USB (Universal Serial Bus) communication is a representative example, and other methods are also possible.

In addition, in the case of the wireless type, a communication method of a Wireless Personal Area Network (WPAN) series such as Bluetooth or Zigbee may be mainly used. However, since the wireless communication protocol is not limited thereto, the wireless communication module may use a wireless local area network (WLAN)-based communication method such as Wi-Fi or other known communication methods.

The first memory 2020 may store various types of information. Various data may be temporarily or semi-permanently stored in the first memory 2020 . Examples of the first memory 2020 include a hard disk (HDD), a solid state drive (SSD), a flash memory, a read-only memory (ROM), and a random access memory (RAM). ), and so on.

The first memory 2020 may be provided in a form embedded in the image analysis apparatus 2000 or in a detachable form. The first memory 2020 includes an operating system (OS) for driving the image analysis apparatus 2000 or a program for operating each component of the image analysis apparatus 2000 , and an operation of the image analysis apparatus 2000 . Various data necessary for the . For example, various data related to images, information related to scan conditions, and information related to analysis results may be stored in the first memory 2020 .

The first controller 2030 may control the overall operation of the image analysis apparatus 2000 . For example, the first controller 2030 may load and execute a program for the operation of the image analysis apparatus 2000 from the first memory 2020 .

The first controller 2030 may be implemented as a central processing unit (CPU) or a similar device according to hardware, software, or a combination thereof. In hardware, it may be provided in the form of an electronic circuit that performs a control function by processing an electrical signal, and in software, it may be provided in the form of a program or code for driving a hardware circuit.

Meanwhile, referring again to FIG. 3 , the image analysis apparatus 2000 according to an embodiment of the present application may include an input module 2040 and an output module 2050 .

In this case, the image analysis apparatus 2000 may obtain a user's input using the input module 2040 and the output module 2050 and output information corresponding to the user's input. For example, the image analysis apparatus 2000 uses the input module 2040 to obtain a user input requesting data acquisition, a user input instructing pre-processing, a user input related to image segmentation, and a reference scan condition related to calculation of morphological indicators A user input may be obtained for , and corresponding information may be output through the output module 2050 .

For example, the user may input conditions or settings related to analysis of the image analysis apparatus 2000 through the input module 2040 .

For example, the user may set a reference scan condition related to a correction parameter for correcting a morphological value or a morphological index obtained from the target image through the input module 2040 . In this case, the image analysis apparatus 2000 may correct a morphological value or a morphological index based on the reference scan condition received from the input module 2040 .

The input module 2040 may be implemented in various forms, such as a mouse, a keyboard, and a touch pad.

The output module 2050 may be provided to output a notification or an image analysis result during an image analysis operation of the image analysis apparatus 2000 .

For example, when the image analysis apparatus 2000 performs an operation of checking the quality of an image, a notification window indicating that an artifact exists in the target image may be provided through the output module 2050 .

As another example, when the image analysis apparatus 2000 performs an operation to check the quality of the image, a notification window may be provided through the output module 2050 to select whether to analyze the image because there is a serious artifact in the target image. have.

As another example, when the image analysis apparatus 2000 performs a segmentation operation on the target image, the segmentation result may be provided through the output module 2050 .

As another example, when the image analysis apparatus 2000 completes the analysis of the target image, the analysis result of the target image may be provided through the output module 2050 .

The output module 2050 may be implemented in various forms, such as a display.

Also, the image analysis apparatus 2000 may further include a user interface for obtaining a user's input through the input module 2040 and outputting information corresponding to the user's input through the output module 2050 .

In FIG. 3 , the image analysis apparatus 2000 according to an embodiment of the present application is illustrated as including an input module 2040 and an output module 2050 , but this is only an example, and the input module 2040 and output The image analysis apparatus 2000 in which the module 2050 is omitted may be provided.

Meanwhile, the image analysis apparatus 2000 according to an embodiment of the present application may be implemented in the form of a server. In this case, the server may be configured to store the medical image transmitted from the image acquisition apparatus 1000 and information related to the medical image. In addition, the server may be implemented to modify or process the medical image transmitted from the image acquisition device 1000 and information related to the medical image.

In addition, the server of the image analysis apparatus 2000 may be implemented separately from the server of the image acquisition apparatus 1000 , but is not limited thereto, and the server of the image analysis apparatus 1000 and the server of the image analysis apparatus 2000 are a single It may be implemented in one form. In other words, the image acquisition apparatus 1000 and the image analysis apparatus 2000 may be implemented as having a common server.

Meanwhile, referring again to FIG. 2 , the image analysis apparatus 2000 according to an embodiment of the present application may be implemented to communicate with the output apparatus 2600 .

The output device 2600 according to an embodiment of the present application may be implemented to receive the analysis result of the image analysis device 2000 and output the image analysis result to the user in a visual form.

As an embodiment, the output device 2600 according to an embodiment of the present application may be implemented to receive information related to image quality from the image analysis device 2000 and output it to the user. For example, the output device 2600 may be implemented to output an index related to image quality or a comment related to reliability of an image analysis result related to image quality.

As an embodiment, the output device 2600 according to an embodiment of the present application may be implemented to receive an analysis result related to a morphological index from the image analysis device 2000 and output it to a user.

For example, the output device 2600 may be implemented to output an analysis result related to a morphological index of an object to a user by comparing and analyzing it with statistical data of a comparison object group. In this case, the statistical data of the comparison subject group and the result of comparative analysis may be processed and outputted as statistical information in the form of any appropriate graph. Also, the output device 2600 may be implemented to provide analysis results related to morphological indices of a plurality of parts of the object together.

As another example, the output device 2600 may be provided to output, in a visual form, information related to where a body part related to the morphological index corresponds while outputting an analysis result related to the morphological index of the object. will be.

As an embodiment, the output device 2600 according to an embodiment of the present application may be implemented to receive the segmentation result of the image from the image analysis device 2000 and output it to the user. For example, the segmentation result of the image analysis apparatus 2000 may be a labeled result in the image. In this case, the image analysis device 2000 or the output device 2600 may visually process the image according to the anatomical structure of the body based on the labeling result, and the output device 2600 may visually display the anatomical structure of the body. It may be implemented to output an image separated by .

As an embodiment, the output device 2600 according to an embodiment of the present application may be implemented to receive a report from the image analysis device 2000 and output it to the user.

Hereinafter, the configuration and operation of the output device 2600 according to an embodiment of the present application will be described with reference to FIG. 4 . 4 is a block diagram of an output device according to an embodiment of the present application.

The output device 2600 according to an embodiment of the present application may include a second communication module 2610 , a second memory 2620 , and a second controller 2630 .

The second communication module 2610 may communicate with any external device including the image acquisition apparatus 1000 and the image analysis apparatus 2000 . In other words, the output device 2600 may transmit and receive images from the image acquisition device 1000 or an image analysis result from the image analysis device 2000 through the second communication module 2610 . Also, the output device 2600 may transmit/receive data to/from any external devices including a repeater through the second communication module 2610 .

For example, the output device 2600, through the second communication module 2610, the image acquired from the image acquisition device 1000, the analysis result related to the quality of the image acquired from the image analysis device 2000 Data related to reliability, data on a segmentation result of an image, data such as a morphological index, and the like may be received. As another example, the output device 2600 transmits or outputs data related to a user input received from an input module 2640 to be described later to the image analysis device 2000 through the second communication module 2610 . Data processed on the device 2600 may be transmitted to the image analysis device 2000 . As another example, the output device 2600 may access the Internet through the second communication module 2610 to upload data related to a user's input and data processed on the output device 2600 .

The second communication module 2610 is largely divided into a wired type and a wireless type. Since the wired type and the wireless type have their respective advantages and disadvantages, in some cases, the image analysis apparatus 2000 may be provided with both the wired type and the wireless type at the same time.

Here, in the case of the wired type, LAN (Local Area Network) or USB (Universal Serial Bus) communication is a representative example, and other methods are also possible.

In addition, in the case of the wireless type, a communication method of a Wireless Personal Area Network (WPAN) series such as Bluetooth (RPuetooth) or Zigbee may be mainly used. However, since the wireless communication protocol is not limited thereto, the wireless communication module may use a wireless local area network (WLAN)-based communication method such as Wi-Fi or other known communication methods.

The second memory 2620 may store various types of information. Various data may be temporarily or semi-permanently stored in the second memory 2620 . Examples of the second memory 2620 include a hard disk (HDD), a solid state drive (SSD), a flash memory, a read-only memory (ROM), and a random access memory (RAM). ), and so on.

The second memory 2620 may be provided in a form embedded in the output device 2600 or in a detachable form. In the second memory 2620 , an operating system (OS) for driving the output device 2600 or a program for operating each configuration of the output device 2600 , as well as various types of operations necessary for the operation of the output device 2600 , are included in the second memory 2620 . Data may be stored. For example, various data related to an image, information related to an analysis result, and user input data may be stored in the second memory 2620 .

The second controller 2630 may control the overall operation of the output device 2600 . For example, the second controller 2630 may load and execute a program for the operation of the output device 2600 from the second memory 2620 .

The second controller 2630 may be implemented as a central processing unit (CPU) or a similar device according to hardware, software, or a combination thereof. In hardware, it may be provided in the form of an electronic circuit that performs a control function by processing an electrical signal, and in software, it may be provided in the form of a program or code for driving a hardware circuit.

Meanwhile, referring again to FIG. 4 , the output device 2600 according to an embodiment of the present application may include an input module 2640 and an output module 2650 .

In this case, the output device 2600 may obtain a user's input using the input module 2640 and the output module 2650 and output information corresponding to the user's input.

For example, the output device 2600 uses the input module 2640 to obtain a user input requesting data acquisition, a user input instructing pre-processing, a user input related to image segmentation, and a reference scan condition related to calculation of morphological indicators and A related user input may be obtained, and corresponding information may be output through the output module 2650 .

In this case, the output device 2600 may further include a user interface for obtaining a user's input through the input module 2640 and outputting information corresponding to the user's input through the output module 2650 .

As an embodiment, the user may input a condition or setting related to the output of the output device 2600 through the input module 2640 , or may select a part of a plurality of output data.

For example, the user may set a reference scan condition related to a correction parameter for correcting a morphological value or a morphological index obtained from the target image through the input module 2640 . In this case, the output device 2600 corresponds to the corrected morphological value or the corrected morphological index based on the reference scan condition (eg, magnetic field strength, manufacturer, setting parameters of the image device, etc.) received from the input module 2640 . It may be implemented to output an analysis result that is

For example, if the user wants to receive an image analysis result according to the changed magnetic field strength by changing the magnetic field strength, a scan condition for the magnetic field strength may be input through the input module 2640 . In this case, the output device 2600 may be implemented to output a corrected image analysis result based on the scan condition for the magnetic field strength input by the user. In this case, the output device 2600 may transmit input data related to the scan condition for the magnetic field strength input by the user to the image analysis device 2000 , and the image analysis device 2000 may transmit the input data based on the received user input. The image analysis result may be corrected based on the obtained correction parameter and transmitted to the output device 2600 . Of course, the output device 2600 may be implemented to receive the correction parameter from the image analysis device 2000, correct the image analysis result, and output it.

In this regard, it will be described later in more detail with reference to FIGS. 66 and 67 .

However, the above-described content is merely an example, and the output device 2600 may be implemented to receive a user input through the input module 2640 in order to output any type of information that the user wants to be provided without being limited thereto. have.

The input module 2640 may be implemented in various forms, such as a mouse, a keyboard, and a touch pad.

The output module 2650 may be provided to output the image analysis result received from the image analysis apparatus 2000 .

For example, the output device 2600 transmits, through the output module 2650 , information on whether an artifact exists in the target image received from the image analysis device 2000 and the reliability of the analysis result as text or arbitrary information. can be output in an appropriate visual form of

As another example, the output device 2600 may output the segmentation result of the target image received from the image analysis device 2000 through the output module 2650 . In this case, the segmentation result may be implemented to divide and output the anatomical structure of the body in a visual graphic form.

As another example, the output device 2600 may output an analysis result of the target image received from the image analysis device 2000 through the output module 2650 . In this case, the analysis result of the target image may be output in the form of a report. In addition, it may be implemented so that the morphological value or morphological index of an anatomical structure included in the analysis result of the target image is output using a statistical technique such as a graph. In addition, morphological values or morphological indicators of anatomical structures included in the analysis result of the target image may be compared with statistical data of a comparison target group and output using a statistical technique such as a graph.

The output module 2650 may be implemented in various forms, such as a display.

As described above, the image analysis device 2000 and the output device 2600 have been described as separate devices. However, this is only an example, and the image analysis device 2000 and the output device 2600 may be implemented as a single device.

The image analysis apparatus 2000 receives the medical image transmitted from the image acquisition apparatus 1000, and performs pre-processing such as alignment of the medical image, normalization of brightness or intensity of the medical image, and noise removal. ) can be done.

Also, the image analysis apparatus 2000 may receive the medical image transmitted from the image acquisition apparatus 1000 and perform segmentation of the medical image. In this case, the segmentation of the medical image according to an embodiment of the present application may be performed using a learned neural network model.

The image analysis apparatus 2000 according to an embodiment of the present application may perform an operation for controlling the quality of a medical image. Accordingly, according to the image analysis apparatus 2000 according to an embodiment of the present application, the reliability of the medical image analysis result may be increased.

The image analysis apparatus 2000 according to an embodiment of the present application may perform an operation of calculating a morphological index based on a medical image obtained from an object. Accordingly, according to the image analysis apparatus 2000 according to an embodiment of the present application, an accurate morphological index may be calculated in relation to the anatomical structure of the body of the object. Through this, there is an advantageous effect of being able to provide an accurate and objective indicator of the subject's disease. In particular, since a morphological index can be calculated by image segmentation without matching a brain image related to an object to a standard brain model, it is possible to provide a diagnostic auxiliary index related to a personalized brain disease.

The image analysis apparatus 2000 according to an embodiment of the present application may perform an operation of correcting a morphological value or a morphological index in consideration of a scan condition or a body part in which a medical image is obtained. Accordingly, according to the image analysis apparatus 2000 according to an embodiment of the present application, an accurate morphological index may be calculated in relation to the anatomical structure of the body of the object. In other words, there may be a slight error in the morphological value calculated according to the scan condition or body part. It is possible to correct the calculated morphological values or morphological indicators. Through this, there is an advantageous effect of being able to provide an accurate and objective indicator of the subject's disease.

The image analysis apparatus 2000 according to an embodiment of the present application gives priority to various image analysis results according to a state of an object or a diagnosis field, and outputs the selected image analysis results based on the priority. can be done Accordingly, according to the image analysis apparatus 2000 according to an embodiment of the present application, user convenience can be improved by selectively providing index information necessary for a user among various index information obtained through analysis of a target medical image.

Hereinafter, some operations performed by an embodiment of the image analysis apparatus 2000 will be described in more detail.

Hereinafter, for convenience of explanation, an embodiment of analyzing images related to a brain will be mainly described. However, it is not limited to the brain and the various embodiments disclosed in the present application may be applied to medical images of any suitable body part other than the brain.

In addition, hereinafter, medical image, brain image, target image (or target image), etc. are used interchangeably, but this is only for convenience of explanation, and medical image, brain image, target image (or target image) are all image analysis It should be construed as referring to an image to be analyzed by the devices 2000 , 3000 , 4000 .

In addition, the following contents may be performed by the image analysis apparatuses 2000 , 3000 , and 4000 according to an embodiment of the present application. Reference numerals designating the image analysis apparatus are merely used to distinguish and explain operations of the image analysis apparatus for convenience of description, and the image analysis apparatus is not limited by the reference numerals.

The image analysis apparatus 2000 according to an embodiment of the present application may acquire a brain image and information related to the brain image.

In more detail, the image analysis apparatus 2000 may acquire a brain image from the image acquisition apparatus 1000 . Also, the image analysis apparatus 2000 may acquire information related to a brain image from the image acquisition apparatus 1000 or an arbitrary external device (eg, a server).

In more detail, the image analysis apparatus 2000 may acquire a brain image and data related to the brain image from the image acquisition apparatus 1000 through the first communication module 2010 .

In this case, the form of the brain image may be a form of various medical images. For example, the brain image may be in DICOM, Neuroimaging Ingormatics Technology Initiative (Nifti), or any suitable format.

The data related to the brain image may include data included in the brain image, data related to a scan condition in which the brain image is obtained, and data about an object of the brain image.

In this case, the data included in the brain image may be data related to pixels or voxels included in the brain image, data related to the orientation of the brain image, and arbitrary metadata structured with respect to the brain image.

In particular, data related to a scan condition in which a brain image, which will be described later, or data about an object of the brain image are structured with respect to the brain image as metadata.

On the other hand, the data related to the scan condition in which the brain image was acquired may be the magnetic field strength of the image acquisition device 1000 from which the brain image was acquired, a setting parameter of the image device of the image acquisition device 1000, or data related to the manufacturer of the imaging device. have.

Since the morphological values obtained from the brain image may be affected according to the scan conditions, the image analysis apparatus 2000 according to an embodiment of the present application obtains data related to the scan conditions and corrects the morphological values. can be performed. This will be described later in detail with reference to FIGS. 58 to 67 .

The data on the object of the brain image may be personal information or medical information about the object of the target image to be analyzed by the image analysis apparatus 2000 . For example, the data on the subject in the brain image may include personal information related to the subject's gender, age, etc., questionnaire information related to brain diseases (eg, dementia, Alzheimer's disease, depression, stroke, etc.) or information related to an underlying disease, etc. It may include various related medical information.

In particular, since gender and age are important variables in brain disease, data related to the subject's gender and age may be a basis for the image analysis apparatus 2000 to determine or obtain a correction parameter. For example, when a target image to be analyzed is obtained from a first object having a first characteristic related to gender and age, the image analysis apparatus 2000 may be configured to correct a morphological value or a morphological index obtained from the target image. The correction parameter may be determined as the correction parameter obtained from the first brain image and the second brain image obtained from the second object having the first characteristic. For example, the image analysis apparatus 2000 may set a correction parameter for correcting a morphological value or a morphological index obtained from the target image by using a correction parameter obtained from a second object having a similar age and the same sex as that of the first object of the target image. can be decided with

Alternatively, data related to the subject's gender and age may be considered to output a relative percentile of the subject's morphological indices according to gender and age using a statistical technique.

Also, the image analysis apparatus 2000 according to an embodiment of the present application may acquire information related to an operation related to brain image analysis.

Specifically, the image analysis apparatus 2000 may acquire information related to a brain-related template for pre-processing or aligning a brain image from an arbitrary external device. In this regard, it will be described later in detail with reference to FIGS. 5, 6 and 48 .

Also, the image analysis apparatus 2000 may acquire information related to a brain atlas, which is a reference for segmenting a brain image, from an arbitrary external device.

For example, the information related to the brain atlas may be information related to the atlas related to the structure of the brain. For example, atlas related to brain structure include Automated Anatomical Labeling (Tzourio-Mazoyer 2002), Desikan-Killiany Atlas (Desikan 2006), Destrieux Atlas (Destrieux 2010), Harvard-Oxford cortical/subcortical atlases (Makris 2006), MICCAI 2012 Multi-Atlas Labeling Workshop and Challenge (Neuromorphometrics), Hammersmith atlas (Hammers 2003, Gousias 2008, Faillenot 2017), HCP MMP 1.0 (Glasser 2016), JuBrain/Juelich histological atlas (Eickhoff 2005) or MarsAtlas (Auzias 2016) have.

As another example, the information related to the brain atlas may be information related to the atlas related to the function of the brain. For example, atlas related to brain function include Mindboggle 101 (Klein 2012), Cortical Area Parcellation from Resting-State Correlations (Gordon 2016), Consensual Atlas of REsting-state Network (CAREN, Doucet 2019), Brainnetome Atlas parcellation (Fan). 2016), Local-Global Parcellation of the Human Cerebral Cortex (Schaefer 2018), Human Motor Area Template (Mayka 2005), Sensorimotor Area Tract Template (Archer 2017), AICHA: An atlas of intrinsic connectivity of homotopic areas (Joliot 2015), Yeo 2011 functional parcellations (Yeo 2011), PrAGMATiC (Huth 2016), fMRI-based random parcellations (Craddock 2011), Voxelwise parcellations (Lead-DBS), SUIT Cerebellar parcellation (Diedrichsen 2006) or Buckner functional cerebellar parcellation (Buckner 2011) can

The image analysis apparatus 2000 may transmit information on the brain atlas related to the above-described brain to the learning apparatus 2200 , and the learning apparatus 2200 is configured for segmentation of a brain image based on the information on the brain atlas. It can be implemented to train a neural network model.

However, the above-described atlas related to the brain structure or information about the atlas related to brain functions is only an example, and the image analysis apparatus 2000 acquires any appropriate brain related atlas information to obtain an image of the learning device 2200 . It can be implemented as a basis for training an artificial neural network model for segmentation.

The image analysis apparatus 2000 may receive a user's input from the input module 2040 of the image analysis apparatus 2000 or the input module 2640 of the output apparatus 2600 .

The image analysis apparatus 2000 may obtain a user input related to a diagnosis target disease. For example, the image analysis apparatus 2000 may obtain a user input related to a brain disease (eg, dementia, depression, stroke, etc.) corresponding to a diagnosis target disease related to the target image.

The image analysis apparatus 2000 may obtain a user input related to a patient. For example, the image analysis apparatus 2000 may obtain a user input related to a patient's gender, age, name, etc. related to the target image.

The image analysis apparatus 2000 may obtain a user input related to image analysis.

For example, the image analysis apparatus 2000 may obtain a user input related to image preprocessing. For example, the image analysis apparatus 2000 may obtain a user input related to image processing, such as correcting the intensity of the target image or removing noise.

For example, the image analysis apparatus 2000 may obtain a user input related to segmentation. For example, the image analysis apparatus 2000 may obtain a user input for correcting labeling data output from a brain atlas and a neural network model considered for segmentation.

For example, the image analysis apparatus 2000 may receive a user input for a scan condition in which a target image is captured through an input module. Information on the scan condition in which the target image is captured may be obtained by being included as metadata of the target image, but may also be obtained by a user input.

For example, the image analysis apparatus 2000 receives, through an input module, a user input for a reference scan condition related to a reference scan condition related to a correction parameter for correcting a morphological value or a morphological index associated with a brain calculated based on a segmentation result. can do. In this case, the image analysis apparatus 2000 may acquire a correction parameter based on a user input for a reference scan condition and correct a morphological value or morphological index related to the brain. In this regard, it will be described later in detail with reference to FIGS. 66 and 67 .

As another example, when outputting an image analysis result, the image analysis apparatus 2000 may receive a user input for selecting some of the output results or giving priority to the user input through the input module. In this case, the image analysis apparatus 2000 may be implemented to selectively output information on an analysis result to be output based on a user input or to output an analysis result reflecting the user's priority. In this regard, it will be described later in detail with reference to FIGS. 68 to 81 .

The data obtained by the image analysis apparatus 2000 may be stored in the first memory 2020 of the image analysis apparatus 2000 or any external device (eg, a server) of the image analysis apparatus 2000 . Also, data obtained by the image analysis apparatus 2000 may be transmitted to the learning apparatus 2200 or the correction parameter obtaining apparatus 2400 . Alternatively, data obtained by the image analysis device 2000 may be transmitted to the output device 2600 or may be transmitted to an arbitrary external device (eg, a server).

The image analysis apparatus 2000 according to an embodiment of the present application may perform image pre-processing. The image analysis apparatus 2000 may perform preprocessing to improve image analysis accuracy. The image analysis apparatus 2000 may be provided to pre-process the image to derive a more accurate segmentation result before performing the image segmentation operation.

For example, the image analysis apparatus 2000 may be provided to convert the format of the image acquired from the image acquisition apparatus 1000 . Specifically, by unifying the format of the images to be analyzed, the neural network model can be trained more stably and accurately. More specifically, it may be more stable and accurate to perform image analysis using an image having the same format as the image used for training the neural network model. Accordingly, the image analysis apparatus 2000 according to an embodiment of the present application may be provided to convert the format of the image acquired from the image acquisition apparatus 1000 .

For example, the image analysis apparatus 2000 according to an embodiment of the present application may perform an operation of converting an image of a first format obtained from the image acquisition apparatus 1000 into an image of a second format. For example, the format of the image acquired from the image acquisition apparatus 1000 may be a DICOM format commonly used in medical images. In this case, the segmentation or morphological index of the image may be calculated based on the DICOM format image.

However, in analyzing a brain image, in the brain image analysis system, it may be relatively easy to analyze an image using a Nifti-type brain image or to train an artificial neural network. Accordingly, the image analysis apparatus 2000 according to an embodiment of the present application may be provided to convert the format of the image acquired from the image acquisition apparatus 1000 into the Nifti format.

However, the above-described format is only an example, and the image analysis apparatus 2000 may be provided to perform a conversion operation into any suitable format other than the Nifti format, if necessary. In addition, the format of the image acquired from the image acquisition device 1000 may be a medical image of any format other than the DICOM format, and even in this case, the image analysis device 2000 is provided to perform a conversion operation into any appropriate format it could be

For example, the image analysis apparatus 2000 may be provided to remove noise that may exist in the image acquired from the image acquisition apparatus 1000 or to correct artifacts. For example, a blurring technique and a technique using a median filter may be used to remove noise. The image analysis apparatus 2000 can derive a more accurate image segmentation result by removing noise and correcting artifacts, and can calculate a morphological index based on the segmentation result with improved accuracy. Therefore, high reliability. This can provide objective diagnostic aids for diseases related to the brain.

As an example, the image analysis apparatus 2000 may be provided to perform an operation of correcting an intensity of an image obtained from the image acquisition apparatus 1000 . By appropriately correcting intensity, noise that may exist in an image can be removed, and an image specialized for an anatomical structure of the brain to be analyzed can be obtained.

As an example, the image analysis apparatus 2000 may be provided to perform an operation of smoothing the image acquired from the image acquisition apparatus 1000 . For example, a technique using blurring (RPurring) or a Gaussian filter (Gaussian filter) may be used as a method of smoothing an image.

As an example, the image analysis apparatus 2000 may be provided to adjust a ratio of the image acquired from the image acquisition apparatus 1000 or perform an operation of cropping the image. For example, the image analysis apparatus 2000 may be implemented to utilize any suitable cropping technique to crop the image. Alternatively, the image analysis apparatus 2000 may adjust the ratio of the image to any suitable image, such as on-demand image resizing, lambda image resizing, a resizing method using a CILanczosScaleTransform filter, or a resizing method using a CIFilter. It may be implemented to utilize a resizing technique.

As an example, the image analysis apparatus 2000 may perform an operation of converting an image obtained from the image obtaining apparatus 1000 into an image captured under a scan condition different from a scan condition in which the image was captured. For example, the image analysis apparatus 2000 may be provided to convert an image captured under the first scan condition of the MRI apparatus into an estimated image as captured under the second scan condition of the MRI apparatus. This image conversion operation may be performed using an MR conversion technique or an artificial intelligence model. However, the present invention is not limited thereto, and the image analysis apparatus 2000 may be provided to perform image conversion in consideration of scanning conditions using any suitable software or image processing technology.

As an example, the image analysis apparatus 2000 may be implemented to perform a pre-processing operation corresponding to a pre-processing operation on an image in the learning apparatus 2200, which will be described later. For example, when the learning device 2200 trains the neural network model using the first pre-processing technique on the image, the image analysis device 2000 pre-processes the target image using the pre-processing technique corresponding to the first pre-processing technique. can be implemented. Through this, image segmentation using a neural network model can be implemented more stably and accurately.

Hereinafter, an image alignment-related operation of the image analysis apparatus 2000 according to an embodiment of the present application will be described with reference to FIGS. 5 to 6 .

5 illustrates an example of an image alignment operation of the image analysis apparatus 2000 .

6 illustrates an example of an image alignment operation of the image analysis apparatus 2000 .

The image analysis apparatus 2000 may be implemented to align the brain images based on data related to the orientation of the brain images included in the brain images.

As an example, the image analysis apparatus 2000 may be implemented to perform an operation of aligning the brain image before performing the image segmentation operation.

For example, referring to FIG. 5 , the image acquisition apparatus 1000 considers the direction RAS of the reference coordinate axis of the object 100 and acquires data (i, j, k) related to the direction of the photographed image together. can do. In detail, the image acquisition apparatus 1000 determines the image obtained by considering the information (x, y, z) on the coordinate axes of the image acquisition apparatus 1000 and the information (RAS) on the reference coordinate axes of the object 100 . Direction-related data (i, j, k) can be obtained.

Data related to the orientation of the image may be structured as metadata for the image. Alternatively, the image may be transmitted to the image analysis apparatus 2000 separately from the image.

The image analysis apparatus 2000 may be implemented to align the images to correspond to the RAS direction (Right-Anterior-Superior direction) of the object 100 based on the data (i-j-k) related to the image direction.

According to the image alignment operation of the image analysis apparatus 2000 described above, before performing the segmentation operation, since brain images, which are the basis of segmentation, can be aligned in a common direction, inaccurate segmentation results are prevented in advance, and the It is possible to secure the stability of the segmentation operation.

The image analysis apparatus 2000 may be implemented to perform spatial normalization of a brain image.

Specifically, an artificial neural network model for segmentation of a brain image may be stably driven with respect to a brain image corresponding to a spatial distribution of a learning image used as "learning data". In other words, if the brain image is different from the spatial distribution of the training image, there is a possibility that the trained artificial neural network may not be driven stably.

Accordingly, the image analysis apparatus 2000 may be implemented to perform spatial normalization of the brain image in order to reduce spatial uncertainty of the brain image.

For example, the image analysis apparatus 2000 according to an embodiment of the present application may be implemented to perform spatial normalization of a brain image based on a brain template. Specifically, the image analysis apparatus 2000 may convert the coordinates of the brain image so that the spatial distribution of the brain image becomes optimal for the artificial neural network model by matching the brain image to the brain template.

For example, a brain template for registering a brain image may be an MNI template, a Talairach template, or the like.

As described above, it has been described that the image analysis apparatus 2000 performs an operation of matching the brain image to the brain template. In this case, "matching" may mean simply matching the positional space of the brain image, rather than an operation corresponding to the internal elements of the brain of the brain template.

According to the spatial normalization operation of the image analysis apparatus 2000 described above, it is possible to remove spatial uncertainty for the image, thereby securing the stability of the segmentation operation using the neural network, and obtaining a segmentation result with improved accuracy.

In addition, data related to the transformed coordinates can be generated by matching the brain image to the brain template. Data related to the transformed coordinates may be used to transform the original coordinates of the brain image after segmentation, which will be described later, is completed. In other words, the image analysis apparatus 2000 may be implemented to convert the segmentation result obtained based on the spatial distribution of the brain template to correspond to the original brain image by using the data related to the transformed coordinates.

Through the operation of the image analysis apparatus 2000, the user can be provided with the analysis result for the brain image with respect to the original coordinates.

For example, when a morphological index related to a brain image is output through an output module, it may be output as a morphological index corrected using a correction parameter. When outputting visual information of the hemiplegia and brain (eg, info. 13 shown in FIG. 57, or the segmentation result, etc.) through the output module, the image is inversely transformed based on the transformed coordinates, that is, the original brain image. Based on the visual information related to the brain may be output.

Hereinafter, operations related to image segmentation of the learning apparatus 2200 and the image analysis apparatus 2000 according to an embodiment of the present application will be described with reference to FIGS. 7 to 17 .

According to an embodiment of the present application, the image segmentation operation may be performed using a learned neural network model. However, even if a neural network model is not used, the image segmentation operation according to an embodiment of the present application may be implemented using any suitable method.

Hereinafter, the operation of learning the neural network model for image segmentation and the content of performing image segmentation using the learned neural network model will be mainly described.

See FIG. 7 . 7 is a diagram illustrating a flowchart of a process for image segmentation according to an embodiment of the present application.

Referring to FIG. 7 , the process for image segmentation according to an embodiment of the present application includes a learning process (P1000) of an artificial neural network model for image segmentation and a segmentation process (P2000) of a target image using the learned artificial neural network model. may include.

In this case, the learning process P1000 may be implemented by the learning apparatus 2200 according to an embodiment of the present application.

Also, the segmentation process P2000 may be implemented by the image analysis apparatus 2000 according to an embodiment of the present application.

In this case, the parameters of the neural network model obtained by the learning process P1000 implemented in the learning apparatus 2200 may be transmitted to the image analysis apparatus 2000 through any suitable communication module.

In this case, the image analysis apparatus 2000 may be implemented to perform segmentation of the target image based on the parameters of the neural network model obtained by the learning process P1000.

The learning process (P1000) according to an embodiment of the present application includes a process (P1100) of acquiring an image data set, a process (P1200) of training a neural network model, a process (P1300) of verifying a neural network model, and parameters of the neural network model and a process P1400 of obtaining.

Hereinafter, a method of learning a neural network model of the learning apparatus 2200 according to an embodiment of the present application will be described with reference to FIG. 8 . 8 is a flowchart of a learning method of a neural network model of the learning apparatus 2200 according to an embodiment of the present application.

Referring to FIG. 8 , the method for learning a neural network model of the learning apparatus 2200 according to an embodiment of the present application includes acquiring an image data set (S1100), screening the image data set (S1200), and image data It may include preprocessing and sorting of the set (S1300), learning and verifying the neural network model (S1400), and obtaining the neural network model parameters (S1500).

In the step of obtaining the image data set ( S1100 ), the learning apparatus 2200 according to an embodiment of the present application may obtain the image data sets from the image obtaining apparatus 1000 or any external devices.

See FIG. 9 . 9 is an exemplary structural diagram of an image data set according to an embodiment of the present application.

The image data set DS obtained by the learning apparatus 2200 may include at least one or more image data. In other words, the image data set DS obtained by the learning device 2200 may include at least one or more image data such as the first image data ID1, the second image data ID2, and the nth image data IDn. can

In this case, the image data included in the image data set DS acquired by the learning apparatus 2200 may include data related to an image and a label.

For example, referring to FIG. 9 , the first image data ID1 included in the image data set DS may include data related to the first image I1 and the first label L1 .

Specifically, the first label L1 may be obtained by manually labeling the first image I1 from a clinician capable of diagnosing a brain disease. Alternatively, the first label L1 may be automatically labeled and obtained using any suitable image segmentation technique.

Data related to images and labels included in image data may be a basis for training an artificial neural network model in relation to a learning method according to an embodiment of the present application and validating the artificial neural network model.

Meanwhile, image data included in the image data set DS may include data related to a scan condition. In this case, the data related to the scan condition may be data related to the magnetic field strength, a setting parameter of the imaging device, and/or data related to the manufacturer of the imaging device, as described above. In addition, data related to the scan condition may be structured as meta data with respect to the image data.

For example, referring to FIG. 9 , the first image data ID1 may include data related to the first scan condition SC1 related to the scan condition in which the first image data ID1 was captured. For example, if the first image data ID1 is photographed under a magnetic field strength of 3T, information related to the magnetic field strength corresponding to 3T is structured as metadata with respect to the first image data ID1 to obtain a data set (S1100) ) through the learning device 2200 .

The data related to the scan condition included in the image data may be considered to obtain a correction parameter for correcting a morphological value or a morphological index corresponding to a target element obtained from a target image, which will be described later.

9 shows only the data included in the first image data ID1, but this is only an example, and the image data of the image data set including the second image data ID2 or the nth image data IDn is It may contain data related to images, labels, and scanning conditions.

Additionally, the learning apparatus 2200 may be implemented to acquire information related to a brain atlas related to a brain structure or a brain function for segmentation related to a brain image.

Specifically, the learning apparatus 2200 may acquire information related to the brain atlas related to the above-described brain structure or brain function from the image analysis apparatus 2000 or any external device.

In this case, the learning apparatus 2200 may be implemented to train an artificial neural network model to be described later in consideration of information related to the brain atlas or to verify the artificial neural network model.

In the step of screening the image data set (S1200), the learning apparatus 2200 according to an embodiment of the present application screens the image data set obtained in the step (S1100) of obtaining the image data set or includes it in the image data set It may be implemented to perform an operation of selecting only some image data from among the image data.

For example, some image data among the acquired image data sets may not be suitable for training an artificial neural network model for segmentation. For example, some image data may include serious artifacts or significant noise. Such image data may not be suitable for training artificial neural network models.

Accordingly, the learning apparatus 2200 may be implemented to screen image data included in the acquired image data set or to select image data effective for training an artificial neural network model.

In the step (S1300) of pre-processing and aligning the image data set, the learning apparatus 2200 according to an embodiment of the present application removes noise or artifacts of an image included in the image data set, or pre-processes to correct the intensity of the image It can be implemented to perform an action.

In addition, the learning apparatus 2200 according to an embodiment of the present application is implemented to align an image based on data related to the orientation of the image, or to align the image by spatial normalization by matching the image to a brain template can be

In this regard, the pre-processing operation of the image analysis apparatus 2000 described above and the image alignment operation described above with reference to FIGS. 5 to 6 may be provided to be implemented in the learning apparatus in the same manner. Alternatively, the image analysis apparatus 2000 may perform image preprocessing and alignment operations through data transmission/reception between the learning apparatus 2200 and the image analysis apparatus 2000 , and then transfer the image to the learning apparatus 2200 .

In the learning and verification of the neural network model ( S1400 ), the training apparatus 2200 for image segmentation according to an embodiment of the present application may train an artificial neural network model for image segmentation.

Specifically, the artificial neural network model may include an input layer for receiving image data, an output layer for outputting a labeling result that is a segmentation result, and a hidden layer including at least one node.

In this case, the learning apparatus 2200 may be implemented to receive image data included in the obtained image data set through an input layer, and obtain a labeling result for the image data obtained by the neural network model through an output layer. .

For example, the learning apparatus 2200 may be implemented to learn an artificial neural network configured to receive the first image data ID1 as an input and output the first 'label L1' through an output layer. Also, the learning apparatus 2200 may input the second image data ID2 as an input layer and obtain a second 'label L2' output through the output layer.

Also, the learning apparatus 2200 according to an embodiment of the present application may learn the artificial neural network in consideration of the brain atlas related to the above-described brain structure or the brain atlas related to brain functions.

The learning apparatus according to an embodiment may learn a neural network model to segment a brain image based on a predetermined brain atlas. The brain atlas may include a plurality of brain regions including a first brain region and a second brain region. The learning apparatus uses an image in which a first region corresponding to the first brain region and a second region corresponding to the second brain region are labeled, and a neural network model to obtain the first region and the second region by inputting the image. can learn

For example, the learning apparatus 2200 according to an embodiment of the present application is implemented to train an artificial neural network model for segmentation of image data included in image data sets based on Desikan-Killiany Atlas (Desikan 2006). can be Desikan-Killiany Atlas (Desikan 2006) is an atlas used to acquire regions corresponding to a plurality of brain regions in a cerebral cortex including a plurality of brain regions including a first brain region and a second brain region.

At this time, the learning apparatus 2400 uses the image data in which the first region corresponding to the first brain region and the second region corresponding to the second brain region are labeled in consideration of Desikan-Killiany Atlas (Desikan 2006), It is possible to train a neural network model to segment cortical regions of the brain.

However, the above-described brain atlas is merely an example, and any suitable brain atlas may be considered according to the purpose of segmentation of image data or a region of interest.

Also, the learning apparatus 2200 according to an embodiment of the present application may be implemented to train a neural network model according to a scan condition in which image data is captured.

For example, the learning apparatus 2200 may be implemented to use the first neural network model to segment a plurality of regions from the first image in the first image obtained under the first scan condition. On the other hand, the learning apparatus 2200 may be implemented to use the second neural network model to segment a plurality of regions from the second image in the second image obtained under the second scan condition. In other words, the learning apparatus 2200 may differently train the neural network model according to a scan condition in which an image is captured.

In addition, in the segmentation process P2000, the image analysis apparatus 2000 according to an embodiment of the present application segments the target image acquired under the first scan condition using the learned first neural network model, and the second scan condition The target image obtained under can be implemented to segment using the learned second neural network model.

Through this, since the image analysis apparatus 2000 according to an embodiment of the present application can perform segmentation of the target image using an optimal neural network model for each scan condition, it is possible to more stably and accurately acquire a plurality of regions. have.

Hereinafter, an example of an artificial neural network model that can be used by the learning apparatus 2200 according to an embodiment of the present application will be described with reference to FIGS. 10 to 11 .

10 is an example of an artificial neural network model that can be used by the learning apparatus 2200 according to an embodiment of the present application.

11 is another example of an artificial neural network model that can be used by the learning apparatus 2200 according to an embodiment of the present application.

Referring to FIG. 10 , a learning apparatus 2200 according to an embodiment of the present application may utilize U-net as an artificial neural network for image segmentation.

U-net used for image segmentation may have an architecture including a contraction path and an expansion path.

Specifically, the contraction path of the U-net may be configured such that two convolutions and max pooling are successively performed. In this case, image-related characteristics may be extracted from the contraction path of the U-net.

However, since the size of the characteristic map is also reduced in the constricted path, the U-net may be configured to recover the size of the characteristic map by additionally including the expansion path.

The extension path of the U-net may be configured such that up-convolution and two convolutions are successively performed. In this case, the size of the image and the characteristic map may be extracted from the extension path of U-net.

Additionally, the U-net architecture is configured to concatenate the characteristic map of the same level, so that it is possible to provide positional information related to the characteristic from the contraction path to the expansion path.

At this time, based on the difference between the label of the input image and the label of the output segmentation map, the parameter of at least one node of the included layer of U-net so that the difference between the label of the input image and the label of the output segmentation map is minimized Alternatively, the weight may be adjusted.

Also, referring to FIG. 11 , the learning apparatus 2200 according to an embodiment of the present application may utilize U-net++ as an artificial neural network for image segmentation. U-net++ is an artificial neural network model using the high-density block idea of DenseNet to improve the performance of U-net. A convolutional layer exists in the skip path to bridge the semantic gap between the encoder and decoder characteristic maps, and to the skip path. It differs from U-net in that dense skip connections exist to improve gradient flow.

Specifically, the learning apparatus 2200 may be implemented to input an input image to the input layer of the U-net++ neural network model and obtain label information output through the output layer. In this case, the learning apparatus 2200 may adjust a parameter or weight of at least one node of the hidden layer included in Unet++ based on a difference between the label information included in the input image and the label information output from the neural network model.

Specifically, the learning apparatus 2200 is implemented to repeatedly perform the operation of adjusting the parameters or weights of at least one node described above, so that the difference between the label information included in the input image and the label information output from the neural network model is minimized. You can obtain the parameters or weights of the node.

As described above, the learning apparatus 2200 according to an embodiment of the present application may perform an operation of learning the artificial neural network model based on the output label result.

Specifically, in the step (S1400) of learning the artificial neural network model, data related to the label included in the image data obtained from the step (S1100) of acquiring the image data set may be acquired.

In this case, the learning apparatus 2200 may be implemented to learn the artificial neural network model based on the image data and the labeling data output through the output layer of the neural network model.

More specifically, the learning apparatus 2200 is based on the difference between the labeling data included in the image data and the labeling data output through the output layer of the neural network model, the weight of at least one node included in the hidden layer of the neural network model or It can be implemented to train the neural network model by adjusting the parameters.

As an example, the learning apparatus 2200 may obtain the labeling data corresponding to the 1A label L1A by inputting the first image data ID1 as an input layer of the artificial neural network. In this case, the learning apparatus may train the neural network model based on the labeling data corresponding to the first label L1 included in the first image data ID1 and the labeling data related to the first label L1A. For example, the learning apparatus 2200 performs a neural network model by adjusting the weights or parameters of at least one node included in the hidden layer of the neural network model based on the difference between the first label L1 and the first label L1A. It can be implemented to learn.

As another example, the learning apparatus 2200 may obtain the labeling data corresponding to the iA-th label LiA by inputting the i-th image data IDi to the input layer of the artificial neural network. In this case, the learning apparatus may train the neural network model based on the labeling data corresponding to the i-th label Li included in the i-th image data IDi and the labeling data related to the iA-th label LiA. For example, the learning apparatus 2200 performs a neural network model by adjusting the weights or parameters of at least one node included in the hidden layer of the neural network model based on the difference between the i-th label (Li) and the iA label (LiA). It can be implemented to learn. Here, i may be any number.

In step S1400 of verifying the artificial neural network model, the learning apparatus 2200 according to an embodiment of the present application may verify the artificial neural network model.

For example, the learning apparatus 2200 according to an embodiment of the present application may obtain labeling data output through a learned neural network model based on at least one image data included in the image data set DS. . In this case, the learning apparatus 2200 may verify the learned neural network model based on labeling data related to at least one image data and labeling data output through the learned neural network model.

For example, the learning apparatus 2200 compares the similarity between labeling data related to at least one image data and labeling data output through the learned neural network model, so that the parameter or weight of the node of the hidden layer of the learned neural network model is You can verify that it is appropriate.

In the step of acquiring the artificial neural network model (S1500), the learning apparatus 2200 according to an embodiment of the present application trains the artificial neural network model on image data included in the image data set and verifies the artificial neural network model By repeatedly performing , it is possible to obtain a neural network model including at least one node having a weight or parameter that minimizes a difference between label-related data included in image data and label-related data output from the artificial neural network.

The obtained node parameters or weights may be used in an artificial neural network model for image segmentation of the segmentation process P2000.

As described above, although segmentation using an artificial neural network has been mainly described, the image analysis apparatus 2000 disclosed in the present application may use various image segmentation algorithms including image segmentation using an artificial neural network.

As an example, the image segmentation algorithm may be provided as a machine learning model. A representative example of the machine learning model may include an artificial neural network. Specifically, a representative example of an artificial neural network is a deep learning-based artificial neural network that includes an input layer that receives data, an output layer that outputs a result, and a hidden layer that processes data between the input and output layers. . Specific examples of artificial neural networks include a convolutional neural network, a recurrent neural network, a deep neural network, a generative adversarial network, and the like, and in the present specification, artificial The neural network should be interpreted in a comprehensive sense including all of the artificial neural networks described above, other various types of artificial neural networks, and artificial neural networks of a combination thereof, and does not necessarily have to be a deep learning series.

In addition, the machine learning model does not necessarily have to be in the form of an artificial neural network model, and in addition, nearest neighbor algorithm (KNN), random forest (RandomForest), support vector machine (SVM), principal component analysis (PCA), etc. may be included. The above-mentioned techniques may include an ensemble form or a form combined in various other ways. On the other hand, it is stated in advance that the artificial neural network can be replaced with another machine learning model unless otherwise specified in the embodiments mentioned mainly with the artificial neural network.

Furthermore, in the present specification, the image segmentation algorithm is not necessarily limited to a machine learning model. That is, the image segmentation algorithm may include various judgment/decision algorithms rather than a machine learning model.

Therefore, in this specification, the image segmentation algorithm should be understood as a comprehensive meaning including all types of algorithms that perform segmentation using image data.

Referring back to FIG. 7 , the segmentation process P2000 according to an embodiment of the present application may include a data acquisition process P2100 and a segmentation process P2200 using a learned neural network model.

The segmentation process P2000 may be implemented by the image analysis apparatus 2000 according to an embodiment of the present application.

Hereinafter, an image segmentation operation using a neural network model of the image analysis apparatus 2000 according to an embodiment of the present application will be described with reference to FIG. 12 . 12 is a flowchart of an image segmentation method using a neural network model of the image analysis apparatus 2000 according to an embodiment of the present application.

Referring to FIG. 12 , the image segmentation method using the neural network model of the image analysis apparatus 2000 according to an embodiment of the present application includes acquiring target image data ( S2100 ), and acquiring segmentation information using a learned neural network. and outputting segmentation information (S2300).

Specifically, in the step of acquiring the target image data ( S2000 ), the image analysis apparatus 2000 may acquire the target image from the image acquiring apparatus 1000 . Also, the image analysis apparatus 2000 may acquire target object information related to the target image or information related to a scan condition in which the target image is captured from the image acquisition device 1000 or any external device.

In this case, target object information related to the target image or information related to a scan condition in which the target image is captured may be considered in acquiring a correction parameter to be described later. In this regard, it will be described in detail with reference to FIGS. 58 to 67 .

See FIG. 13 . 13 is an exemplary structural diagram of a target image according to an embodiment of the present application.

For example, target image data acquired by the image analysis apparatus 2000 according to an embodiment of the present application may include information on the target image TI. For example, the information on the target image TI may include information related to coordinates, intensity, color, etc. of a pixel.

As another example, the target image data may include target object information TO. For example, the information on the target object information TO may be information on personal information of an object (eg, a brain disease test patient) related to the target image TI. For example, the information on the target object information TO may be information related to a name, age, gender, etc. of an object (eg, a brain disease test patient). In this case, the image analysis apparatus 2000 may acquire information on the target object information TO from any external device. Alternatively, the image analysis apparatus 2000 may acquire information on the target object information TO by recognizing the metadata structured with respect to the target image data.

As another example, the target image data may include information about the target scan condition (TSC). For example, the information on the target scan condition TSC may be information related to the scan condition in which the target image TI is captured. For example, information on the target scan condition TSC may include a magnetic field strength in which the target image TI is photographed, a setting parameter of an imaging device in which the target image TI is photographed, or an image device in which the target image TI is photographed. It may be information related to the manufacturer. In this case, the image analysis apparatus 2000 may acquire information on the target scan condition TSC from any external device. Alternatively, the image analysis apparatus 2000 may acquire information on the target scan condition (TSC) by acquiring structured metadata with respect to the target image data.

Referring back to FIG. 7 , the image analysis apparatus 2000 may be implemented to input the target image data acquired in the data acquisition process P2100 as an input layer of the learned neural network model.

At this time, the image analysis apparatus 2000 is configured to segment the target image data using the node weights and/or node parameters of the artificial neural network model obtained according to the learning process P1000 implemented in the learning apparatus 2200 described above. It can be implemented to be used in an artificial neural network model.

Again, referring to FIG. 12 , in the step of obtaining segmentation information using the learned neural network model ( S2200 ), the learned neural network model based on the node weight and/or the node parameter obtained from the learning device 2200 is , may be provided to receive target image data through an input layer and output a result of labeling the target image TI as a result of segmenting the target image TI through the output layer. In this case, the image analysis apparatus 2000 may acquire segmentation information related to the labeling of the target image through the output layer of the learned neural network model.

In this case, the result output through the output layer of the artificial neural network model may include a plurality of target regions obtained from the target image of the target image data.

A result output through the output layer of the neural network model may be in a labeling form corresponding to a plurality of target regions obtained from the target image. For example, a result output through the output layer of the neural network model may be in the form of labeling data including a first label defining a first region and a second label defining a second region obtained from the target image.

In this case, the result output through the output layer of the neural network model may be in a form in which a first color is overlaid on a first area of the target image and a second color is overlaid on a second area of the target image based on the first label. Through this, the first area and the second area may be more easily distinguished. However, the above description is only an example, and the output result may be configured in an arbitrary form for distinguishing the first area and the second area.

Also, the segmentation information output through the output layer of the neural network model may be in the form of dividing the target image into a plurality of regions based on a predetermined brain atlas. For example, as described above, the learning apparatus trains the neural network model to divide the learning image into a first region corresponding to the first brain region and a second region corresponding to the second brain region based on a predetermined brain atlas. can At this time, since a neural network model learned based on a predetermined brain atlas is used, the segmentation information output through the output layer of the neural network model is applied to the first and second brain regions corresponding to the first brain region of the target image. The output may be divided into a plurality of regions including two regions.

The image analysis apparatus 2000 may perform an operation of calculating a morphological value corresponding to a specific region based on the segmentation information output through the output layer.

As an example, the image analysis apparatus 2000 may be implemented to calculate a morphological value representing a morphological character related to the first brain region based on the segmentation information related to the first region output through the output layer. have. For example, a morphological character may relate to volume, thickness, length or shape, and the like.

In the step of outputting the segmentation information ( S2300 ), the image analysis apparatus 2000 overlays visual graphics on a plurality of brain regions based on the segmentation information output through the output layer to the output module of the output apparatus 2600 . It may be implemented to be displayed to the user through the output module 2050 of the 2650 or the image analysis apparatus 2000 .

See FIG. 14 . 14 is an example of images obtained based on segmentation information obtained by a segmentation process of the image analysis apparatus 2000 according to an embodiment of the present application.

Specifically, the diagram shown in the upper part of FIG. 14 is an example of an image output based on a segmentation result for an image obtained from T1-MRI. 14 is an example of an image output based on a segmentation result for an image obtained from T2-Flair MRI.

For example, in the image analysis apparatus 2000 according to an embodiment of the present application, the frontal lobe, temporal lobe, parietal lobe, occipital lobe, lateral ventricle, amygdala, hippocampal region, etc. Corresponding segmentation information may be obtained.

For example, the image analysis apparatus 2000 according to an embodiment of the present application corresponds to white matter, gray matter, ventricle, white matter hyperintensity (WMH) region, etc. by a segmentation process for an image obtained from T2-Flair MRI. segmentation information can be obtained.

According to the image segmentation operation of the image analysis apparatus 2000 described above, the user can visually check the segmentation result, and thus can easily check the segmentation result. In addition, an advantageous effect of improving the user's understanding of an auxiliary indicator for diagnosing a brain disease may be provided.

Meanwhile, the artificial neural network model used in FIG. 7 may be implemented as at least one artificial neural network model.

Hereinafter, a flowchart of a process for image segmentation using at least one artificial neural network model according to an embodiment of the present application will be described with reference to FIGS. 15 to 17 . Specifically, in relation to the process ( P1200 ) of training the artificial neural network model of FIG. 7 , characteristic contents in the case of using a plurality of artificial neural network models will be mainly described. The contents described in relation to FIGS. 7 to 14 may be analogically applied to the embodiments to be described later in relation to FIGS. 15 to 17 .

15 is a diagram illustrating a flowchart of a process for image segmentation according to the present embodiment.

As an example, referring to FIG. 15 , the neural network model learning process P1200 of the learning process P1000 of FIG. 7 is a process P1210 for learning the first neural network model and a process P1220 for learning the second neural network model. can be configured.

As an example, the first neural network model may be trained to acquire a first region corresponding to the first brain region and a second region corresponding to the second brain region. In this case, the second neural network model may be trained to acquire the third region and the fourth region included in the first region that may be obtained from the first neural network model. In other words, the first neural network model may be trained to acquire regions (eg, cranial region, cerebrospinal fluid (CSF) region, cortical region, medulla region) corresponding to the macroscopic structure of the brain, and the second neural network model may be It may be learned to acquire regions corresponding to the detailed structure of the brain (eg, regions corresponding to brain elements located in the cortex, regions corresponding to brain elements located in the medulla).

As another example, the first neural network model may be trained to obtain regions corresponding to brain regions corresponding to the first brain atlas. On the other hand, the second neural network model may be trained to acquire regions corresponding to brain regions corresponding to the second brain atlas.

As another example, the first neural network model may be trained to acquire regions (eg, cranial region, cerebrospinal fluid (CSF) region, cortical region, and medulla region) corresponding to the macroscopic structure of the brain. On the other hand, the second neural network model may be trained to acquire regions corresponding to brain regions corresponding to the brain atlas. For example, the second neural network model may be trained to acquire regions corresponding to a plurality of brain regions based on Desikan-Killiany Atlas (Desikan 2006), and the second neural network model learned based on Desikan-Killiany Atlas (Desikan 2006) The two neural network model can be trained to acquire regions including the frontal lobe, temporal lobe, parietal lobe, occipital lobe, lateral ventricle, amygdala and hippocampus.

The process P1210 for training the first neural network model and the process P1220 for training the second neural network model may be independently performed.

Specifically, the data sets used in the process P1210 for training the first neural network model and the process P1220 for training the second neural network model may be independent.

See FIG. 16 . 16 is an exemplary structural diagram of image data sets according to an embodiment of the present application. Specifically, the image data set shown on the left side of FIG. 16 may be an image data set for training the first neural network model. On the other hand, the image data set shown on the right side of FIG. 16 may be an image data set for training the second neural network model.

For example, the image data set DS for training the first neural network model includes first image data including information related to a first image I1 and a first label L1a related to first regions. can do. On the other hand, the image data set DS for training the second neural network model may include the first image data including information related to the first image I1 and the 1b label L1b related to the second regions. .

In this case, the learning apparatus uses the first neural network model to input the first image data ID1 as an input layer of the artificial neural network to obtain output labeling data corresponding to the first label L1a' related to the first regions. can do. In this case, the learning apparatus performs labeling data corresponding to the label L1a related to the first regions included in the first image data ID1 and the output labeling data related to the label 1a' label L1a' related to the first region. Based on this, an artificial neural network model can be trained. For example, the learning apparatus 2200 is implemented to adjust the weight or parameter of at least one node included in the hidden layer of the neural network model based on the difference between the 1a label L1a and the 1a' label L1a'. can be Also, by repeatedly performing the above-described process of training the first neural network model, parameters related to the first neural network model related to the first region may be acquired.

On the other hand, the learning apparatus may obtain output labeling data corresponding to the first label L1b' related to the second regions by inputting the first image data to the input layer of the artificial neural network using the second neural network model. have. In this case, the learning apparatus generates labeling data corresponding to the 1b label L1b related to the second regions included in the first image data ID1 and the output labeling data related to the 1b label L1b′ related to the second region. An artificial neural network model can be trained based on For example, the learning device 2200 is implemented to adjust the weight or parameter of at least one node included in the hidden layer of the neural network model based on the difference between the 1b label (L1b) and the 1b' label (L1b'). can be In addition, by repeatedly performing the above-described process of training the second neural network model, parameters related to the second neural network model related to the second region may be acquired.

In addition, the parameters related to the first neural network model obtained in the learning process P1000 may be used in the first neural network model for image segmentation corresponding to the first region of the segmentation process P2000 (P2210), and the learning process (P2210) The parameters related to the second neural network model obtained in P1000) may be implemented to be used (P2220) by the second neural network model for image segmentation corresponding to the second region of the segmentation process (P2000).

However, according to FIG. 16 , the label 1a included in the image data related to the first neural network model is a label related to the first region, and the label 1b included in the image data related to the second neural network model is a label related to the second region. Although shown as being, this is merely an example and is not limited thereto. For example, the image data of the first and second neural network models may include label information related to the first region and the second region, and when training the first neural network model, only label information related to the first region is used and When training the second neural network model, it may be implemented to use only label information related to the second region.

Also, as described above, although it has been described that the first neural network model and the second neural network model are independent, the present invention is not limited thereto, and the first neural network model and the second neural network model may share at least some layers. In other words, the first neural network model and the second neural network model may include at least one common layer.

Meanwhile, the process of learning the first neural network model and the process of learning the second neural network model may be independent, but may be taught to be related to each other. Here, learning to be related to each other includes using data output from one neural network model as input data of the other neural network model, and using arbitrary data generated from one neural network model in another neural network model. It can mean encompassing all forms.

See FIG. 17 . 17 is a diagram illustrating a flowchart of a process for image segmentation according to an embodiment of the present application.

In one embodiment, the process (P1200) of training the neural network model of the learning process (P1000) according to an embodiment of the present application includes a process (P1211) of training a first neural network model and a process (P1211) of training a second neural network model ( P1221), but may be implemented such that data related to a result output from the first neural network model is input to the second neural network model as input data.

As an example, the learning apparatus 2200 may configure the second neural network model to receive labeling data related to the first region. In this case, the labeling data related to the first region may be obtained by manually labeling the first region corresponding to the first brain region with respect to the image data, or may be obtained using any automatic labeling software.

For example, the labeling data related to the first region may be data manually labeled by a clinician with respect to image data input to the first neural network model.

As another example, the first neural network model may be trained to output labeling data related to the first region, and in this case, the labeling data related to the first region input to the second neural network model is the labeling output from the first neural network model. It can be data.

In this case, the second neural network model may be trained to output labeling data related to the second region based on the labeling data and image data related to the first region.

Therefore, the learning apparatus 2200 according to an embodiment of the present application may obtain labeling data related to the first region through the first neural network model and obtain labeling data related to the second region through the second neural network model. can be implemented. Accordingly, the learning apparatus 2200 may train the first neural network model and the second neural network model to obtain labeling data related to the first region and the second region.

In addition, the learning apparatus 2200 is configured to perform at least the first neural network model based on the difference between the labeling data associated with the label associated with the first region and the second region and the labeling data associated with the first region and the second region of the image data set. It may be provided to adjust a weight or parameter of one node or adjust a weight or parameter of at least one node of the second neural network model.

In addition, the parameters related to the first neural network model and the second neural network model obtained in the learning processor P1000 are the first neural network model and the second neural network model for image segmentation corresponding to the first region and the second region of the segmentation process P2000. It can be used for neural network models.

In this case, the first region or the second region may be a region related to an anatomical structure of the brain, and may be one of regions partitioned based on the brain atlas described above. Also, the first region or the second region may be a region showing a significant association related to a brain disease.

However, the above-described learning process of the neural network model is only an example, and the image segmentation operation is learned using various combinations or connections of at least one or more neural network models having any appropriate form, type, and parameter, and the neural network model is verified. It may be possible to obtain the parameters of the neural network model.

Meanwhile, the image analysis apparatus 2000 according to an embodiment of the present application may be implemented to update or update the learned artificial neural network model in the segmentation process P2000.

As an example, the image analysis apparatus 2000 may be implemented such that segmentation information obtained by segmenting the target image by the segmentation process P2200 using the learned neural network model can be modified manually or using any software. . At this time, the image analysis apparatus 2000 may be implemented to update or update the artificial neural network model by modifying the weights of at least one node or the parameters of at least one node of the learned neural network model based on the modified segmentation information. have.

The image analysis apparatus according to an embodiment may perform a function of determining the quality of an image. The image analysis apparatus may have a quality criterion for determining the quality of the image, and may determine whether an input image satisfies the quality criterion.

More specifically, the image quality determination may mean determining whether the medical image acquired by the image acquisition device has a quality above a certain level. In other words, the image quality determination may refer to determining whether medical information having a certain level of reliability or higher can be obtained from the medical image acquired by the image acquisition device.

Image quality determination may be performed together with detailed operations constituting an image analysis operation. For example, the image quality determination operation may be performed together with the image acquisition operation, image preprocessing operation, image segmentation operation, and/or medical information output operation described herein.

The image quality determination operation may be performed before or after the at least one detailed operation is performed. The image quality determination operation may be performed based on information obtained as a result of performing the detailed operation. The image quality determination operation may be performed to determine whether a criterion for performing a detailed operation is satisfied.

The image quality determination may be performed based on image data. The image quality determination may be performed based on the raw image data. The image quality determination may be performed based on a medical image on which pre-processing has been performed. As another example, the image quality determination may be performed based on a result of segmentation of the medical image. As another example, the image quality determination may be performed based on a result of analyzing the medical image.

The image quality determination may be performed based on non-image data related to the image data. The image quality determination may be performed based on at least one of metadata information about the medical image, artifact information included in the medical image, or ROI information secured through medical image segmentation.

According to an embodiment, the image quality determination may be performed differently for each corresponding detailed operation.

For example, the first quality determination may be performed based on the raw data before the preprocessing operation. In this case, a pre-processing operation may be performed on an image that satisfies the first quality criterion. Also, for example, the second quality determination may be performed based on raw data or a pre-processed image before the segmentation operation. In this case, a segmentation operation may be performed on an image that satisfies the second quality criterion. Also, for example, the third quality determination may be performed based on raw data, a preprocessed image, or an image segmentation result before image analysis. Image analysis may be performed on an image that satisfies the third quality criterion. The first to third quality criteria may be different from each other.

Information obtained through image quality determination may be output. The information obtained through the image quality determination may include information that is a basis for the image quality determination. The information that is the basis for determining the image quality may include whether there is a defect in the image, information about the format of the image, conditions for acquiring the image, and the like. Information obtained through image quality determination may be used to generate or provide information related to whether a subsequent operation is performed.

A specific embodiment of image quality determination will be described later.

The image analysis apparatus 2000 according to an embodiment of the present application may be provided to perform an operation of calculating the brain morphology index of the object based on the segmentation result of the object image.

Specifically, the image analysis apparatus 2000 may acquire an inner region of the skull and a region corresponding to the target element based on the segmentation result. Also, the image analysis apparatus 2000 may additionally perform an operation of correcting a boundary corresponding to the inner region of the skull in order to calculate the morphological index of the target element.

Also, the image analysis apparatus 2000 may additionally perform an operation of aligning the brain image in order to correct the boundary of the inner region of the skull.

The operation of calculating the brain morphological index of the object by the image analysis apparatus 2000 according to an embodiment of the present application will be described later in detail with reference to FIGS. 48 to 57 .

The image analysis apparatus 2000 according to an embodiment of the present application may be provided to perform an operation of correcting a brain morphological value or a morphological index calculated based on the segmentation result of the target image.

Specifically, the image analysis apparatus 2000 may acquire a region corresponding to the target element and pixel or voxel data corresponding to the target element based on the segmentation result. In this case, the image analysis apparatus 2000 may acquire the morphological value of the target element based on pixel or voxel data corresponding to the target element.

In addition, the image analysis apparatus 2000 may perform an operation of correcting the morphological value of the target element in consideration of the scan condition in which the target image is obtained or the location of the target element in the target image so as to output a more accurate morphological index. have.

In this case, the image analysis apparatus 2000 may obtain a correction parameter in consideration of the scan condition or the location of the target element from the correction parameter acquisition apparatus 2400 in order to correct the morphological value of the target element, and the correction parameter and the target element It may be provided to perform an operation of outputting a morphological index of the target element based on the morphological value.

The operation of correcting the morphological value or morphological index of the brain of the image analysis apparatus 2000 according to an embodiment of the present application will be described later in detail with reference to FIGS. 58 to 67 .

The image output apparatus according to an exemplary embodiment may provide the user with diagnosis auxiliary information based on various medical information obtained through image analysis. Here, the diagnosis auxiliary information may include information obtained through processing of various types of medical information obtained from a medical image. For example, the diagnosis auxiliary information may include information obtained by processing medical information, for example, diagnosis information based on medical information, analysis information, prescription information, and the like.

The operation of providing diagnostic auxiliary information may be performed together with a detailed operation constituting the operation of outputting an image. For example, the operation of providing diagnostic auxiliary information may include an image acquisition operation described herein, an operation for acquiring medical information from an image, an operation for acquiring diagnostic assistance information among the acquired information, an operation for outputting diagnostic assistance information, and/or a diagnosis assistance operation. It may be performed together with an operation of providing a comment based on the information. A detailed description of each of the above-described operations will be provided later.

The image output apparatus according to another exemplary embodiment may selectively provide index information necessary for a user from among various medical information obtained through image analysis. Here, the selective information provision may include selectively providing only necessary medical information to the user from among various medical information obtainable through the image analysis apparatus.

The optional information providing operation may be performed together with a detailed operation constituting the image output operation. For example, the selective information providing operation may be performed based on the image acquisition operation described herein, the operation of acquiring medical information from the image, the operation of acquiring selective information among the acquired information, the operation of outputting the optional information and/or the selective information. This may be performed in conjunction with the action of providing a comment. A detailed description of each of the above-described operations will be provided later.

In the above, the configuration and operation of the image analysis apparatus 2000 according to an embodiment of the present application have been described. Hereinafter, the image analysis method according to the present embodiment will be described in more detail.

In the following description, it is described that the image analysis method according to an embodiment of the present application is performed by the above-described image analysis apparatus 2000 , the learning apparatus 2200 , the correction parameter obtaining apparatus 2400 or the output apparatus 2600 . do. However, since this is only for convenience of explanation, the image analysis method according to an embodiment of the present application includes the image analysis apparatus 2000, the learning apparatus 2200, the correction parameter obtaining apparatus 2400, or the output apparatus ( 2600) is not limited. That is, the image analysis method described later does not necessarily have to be performed only by the image analysis apparatus 2000 , the learning apparatus 2200 , the correction parameter obtaining apparatus 2400 , or the output apparatus 2600 , and the above-described image analysis apparatus 2000 . , it is also possible to be performed by another system or device having a function similar to that of the learning device 2200 , the correction parameter acquiring device 2400 , or the output device 2600 .

A variety of information is obtained from medical images to determine a patient's health status. In order to obtain accurate information, analysis must be performed based on medical images that have a quality above a certain level and satisfy necessary conditions.

More specifically, the medical image acquired by the image acquisition device may have an unsuitable quality (or property) for image analysis to take place. For example, the image may have formal or substantial imperfections that are inappropriate for medical data acquisition. For example, a medical image may not be consistent due to the brightness of the image, the resolution of the image, the area in which the image was taken, the direction in which the image was taken, the angle at which the image was taken, or various defects that may occur in the image due to other problems that may occur during the image. It may not be of higher quality than the standard. Alternatively, the format of the image, such as the format and size of the image file, may not satisfy the necessary conditions.

When the analysis is performed based on a medical image that does not have a quality above a certain level as described above, the analysis result may also not have reliability above a certain level. Accordingly, in order to derive an analysis result capable of having a higher level of reliability, an image quality determination must be made as to whether an image acquired by the image acquisition device has a quality above a certain level.

In the past, in order to determine whether there are any defects in the medical images acquired through the imaging equipment, the inspector personally judged them one by one. However, there was a limitation in that it was difficult to make a quality judgment with a certain level of accuracy and high quality because the judgment result may not be constant depending on the inspector's point of view, the inspector's experience, and the inspector's condition.

According to an embodiment, artificial intelligence may be utilized to overcome the above-described limitations and to determine the improved quality of a photographed medical image. That is, the image analysis apparatus according to an embodiment may perform high-level medical image quality determination using artificial intelligence.

18 is a diagram for describing an image quality determination process according to an embodiment.

Referring to FIG. 18 , the image quality determination process according to an exemplary embodiment may include a medical image acquisition step S3000 , a medical image quality determination step S3300 , and a medical image quality related information output step S3500 . In this case, the step of outputting medical image quality related information ( S3500 ) may be omitted. Meanwhile, the image quality determination process may further include pre-processing the medical image and/or segmenting the medical image.

19 is a diagram for explaining an image analysis apparatus 3000 that performs an image quality determination process.

Referring to FIG. 19 , the image analysis apparatus 3000 may perform an image quality determination process. The image analysis apparatus 3000 may include one or more modules for determining image quality. For example, the image analysis apparatus 3000 may include at least one of a preprocessing performing module 3300 , an image segmentation module 3500 , an image quality determining module 3700 , and an image quality related information output module 3900 .

Hereinafter, each step of the image quality determination process will be described with more specific examples. The image quality determination process described below may be performed by the above-described image analysis apparatus 3000 or the apparatus or system described throughout this specification.

The image acquisition apparatus according to an embodiment may acquire a medical image. For example, the medical image may include, but is not limited to, images such as computed tomography (CT), magnetic resonance imaging (MRI), and X-ray. In this case, the MRI image may include various types of images that can be captured by MRI imaging equipment, for example, a T1-weighted image, a T2-weighted image, or a FLAIR image. Also, the MRI image may include images photographed on various types of planes, for example, axial, sagittal, or coronal planes photographed by MRI imaging equipment.

In the image acquisition apparatus according to an embodiment, the image analysis apparatus 3000 may acquire a medical image in which at least a portion is not completely photographed. For example, the image analysis apparatus 3000 may acquire a medical image in which at least a portion of the target area is omitted or improperly captured.

For example, the image acquisition apparatus may acquire a medical image having a quality below a certain level. The image acquisition apparatus may acquire a medical image in which it is difficult to expect extraction of medical information having a reliability level above a certain level. As another example, the image acquisition device may acquire a medical image having an abnormal file structure or missing patient information. As another example, the image acquisition apparatus may acquire a medical image including at least one noise. Here, the noise may mean various types of defects or states that affect the acquisition of information based on an image.

The image quality determination process according to an embodiment may include pre-processing the medical image. For example, the image quality determination process according to an embodiment may include preprocessing the medical image acquired by the image acquisition device to be suitable for image quality determination to be performed.

Illustratively, the image pre-processing step may include performing various pre-processing processes, such as correcting the brightness, size, ratio, direction, or resolution of the image so that the detection of artifacts included in the medical image can be performed more smoothly. . As another example, the image pre-processing step includes performing various pre-processing processes such as correcting the brightness, size, ratio, direction, or resolution of the image so that information on anatomical structures included in the medical image can be acquired more smoothly. may include

In addition, the image pre-processing step may include performing various pre-processing processes for determining image quality, and since detailed information on this has been described above, redundant content will be omitted.

An image quality determination process according to an embodiment may include segmenting the medical image. For example, the image quality determination process according to an embodiment may include segmenting the medical image acquired by the image acquisition device for image quality determination.

Exemplarily, the image segmentation step may include performing medical image segmentation to obtain information related to an artifact included in the medical image. As a more specific example, the image segmentation step may include performing segmentation of a region corresponding to an artifact included in the medical image.

The segmentation of the region corresponding to the artifact may be performed using a neural network model trained to obtain the region corresponding to the artifact included in the medical image. Segmenting the artifact region may be performed using a neural network model trained using training data including one or more medical images labeled with the artifact region.

As another example, the image segmentation step may include segmenting the medical image so that image quality can be determined based on at least a part of a region of the human body included in the medical image. As a more specific example, the image segmentation step may include segmenting the medical image to obtain information about an anatomical or functional structure of a human body that may be a basis for determining image quality. The image segmentation step may include segmenting an area corresponding to the structure of the human body included in the medical image. The image segmentation step may include acquiring a region corresponding to a structure used for image quality determination.

Segmentation of a region corresponding to the structure of the human body may be performed using a neural network model. The segmentation of the region corresponding to the structure may be performed using a neural network model trained to segment the region included in the medical image. The image segmentation step may include performing segmentation of the medical image using a pre-trained neural network model to obtain at least one segmented region. In this case, the at least one segmented region may correspond to different anatomical or functional structures, respectively. Also, the at least one segmented region may include a region corresponding to a human body structure used for quality determination. Segmentation of a region corresponding to an anatomical or functional structure may be similarly applied to content related to segmentation of an image described throughout this specification.

20 is a diagram for explaining the image quality determination module 3700 .

Referring to FIG. 20 , the image quality determination module 3700 includes a first image quality determination unit 3710 , a second image quality determination unit 3730 , a third image quality determination unit 3750 , or a fourth image quality determination unit At least one of (3770) may be included.

For example, the first image quality determining unit 3710 may determine the quality of the image based on the metadata information, and the second image quality determining unit 3730 may determine the image quality based on the noise information. The third image quality determining unit 3750 may determine the image quality based on the anatomically segmented segmentation information, and the fourth image quality determining unit 3770 may provide complex information, for example, noise information and anatomy. The quality of the image may be determined based on the relationship between the segmented segmentation information. Details of the first image quality determining unit 3710 to the fourth image quality determining unit 3770 will be described later.

A neural network model for determining image quality according to an embodiment may be differently learned and performed according to a type of a medical image acquired by the image acquisition device. For example, when the type of the medical image acquired by the image acquisition apparatus is a CT image, the image quality determination model of the image analysis apparatus 3000 may be a model that is learned and performed based on the CT image. As another example, when the type of the medical image acquired by the image acquisition apparatus is MRI, the image quality determination model of the image analysis apparatus 3000 may be a model that is learned and performed based on the MRI image.

The first image quality determining unit 3710 according to an embodiment may be performed based on metadata information to determine whether image analysis can be normally performed. In this case, the metadata information may include information about the medical image acquired from the image acquisition device. More specifically, the metadata information may include at least one of file structure information of the medical image and patient information input to the medical image.

21 is a diagram for explaining a first image quality determination process.

Referring to FIG. 21 , the first image quality determination process includes a medical image acquisition step (S3711), a non-image information acquisition step (S3713), a non-image information normal determination step (S3715), and a non-image-related information output step (S3717) may be included.

The non-image information acquisition step S3713 may include acquiring medical image information from a medical image. The non-image information obtaining step S3713 may include obtaining non-image information from the medical image. Here, the non-image information may include file structure information of the medical image or patient information input to the medical image.

In this case, the file structure information of the medical image may include information about the file format of the medical image, the format of the medical image, or the size of the file, but is not limited thereto. In addition, the patient information may include, but is not limited to, information related to personal information such as the patient's name and age, information related to a time when the patient took a medical image, or information about the patient's health condition.

The step of determining whether non-image information is normal ( S3715 ) may include determining image quality based on non-image information extracted from the medical image. The determining whether non-image information is normal ( S3715 ) may include determining image quality based on at least one of file structure information and patient information from a medical image.

According to an embodiment, the determining whether the non-image information is normal ( S3715 ) may determine whether the file structure of the medical image is abnormal. Here, the file structure of the medical image may mean the file format of the medical image, the format of the medical image, or the size of the file, but is not limited thereto.

For example, since the file format or format of the medical image must be a file format or format suitable for image analysis performed by the image analysis apparatus 3000 , the non-image information normal determination step S3715 is a file of the medical image. It may include determining whether the format or format is a file format or format suitable for performing image analysis by the image analysis apparatus 3000 . Accordingly, the determining whether non-image information is normal ( S3715 ) may include acquiring information on whether image analysis can be normally performed based on the file structure information of the medical image.

According to another embodiment, the step of determining whether non-image information is normal ( S3715 ) may include determining whether patient information is missing from the medical image acquired from the image acquisition device. Here, the patient information may include, but is not limited to, personal information such as the patient's name and age, or information about the patient's health status.

For example, if patient information is not included in the medical image, it is not possible to know which patient the image analysis result by the image analysis apparatus 3000 relates to, so the non-image information normal determination step S3715 is It may include determining whether patient information is included. Accordingly, the step of determining whether the non-image information is normal ( S3715 ) may include acquiring information on whether patient information is input to the medical image.

22 is a diagram for explaining the first image quality determining unit 3710 .

Referring to FIG. 22 , the first image quality determining unit 3710 may include at least one of a non-image information obtaining unit 3711 and a non-image information normal determining unit 3717 .

The non-image information obtaining unit 3711 may include a file structure information obtaining unit 3713 or a patient information obtaining unit 3715 . The non-image information obtaining unit 3711 may obtain non-image information based on the obtained medical image. In this case, the acquired medical image may include raw data before the pre-processing operation or an image on which the pre-processing operation is performed.

The non-image information determining unit 3717 may determine whether the non-image information is normal based on the non-image information acquired by the non-image information obtaining unit 3711 . The non-image information normal determination unit 3717 may output information about a result of determining whether the non-image information is normal. For example, the non-image information normal determination unit 3717 may include information on whether image analysis can be normally performed based on the file structure information of the medical image or information on whether patient information is input to the medical image. can be printed out.

The second image quality determining unit 3730 according to an embodiment may perform a function of extracting noise information included in the image. Here, the noise information may mean various types of defects or states that affect the acquisition of information based on an image. For example, the noise information may include information related to the resolution of the image, information related to the brightness of the image, or information about artifacts generated in the image. .

23 is a diagram for explaining a second image quality determination process.

Referring to FIG. 23 , the second image quality determination process includes obtaining a medical image ( S3731 ), determining whether an artifact occurs in the medical image ( S3733 ), obtaining artifact information from the medical image ( S3735 ), and outputting artifact information (S3737) may be included.

The step of determining whether an artifact is generated in the medical image ( S3733 ) may include determining whether an artifact is included in the medical image. Here, the artifact may include noise related to pixels that do not faithfully show anatomical structures in the medical image. Illustratively, the artifact may include a motion artifact, a bias artifact, a zipper artifact, a ghost artifact, a spike artifact, and the like, but is not limited thereto, and may include a variety of known artifacts.

The step of acquiring artifact information from the medical image ( S3735 ) may include acquiring information about the artifact generated in the medical image. Here, the artifact information may include information related to the artifact, for example, information on whether or not the artifact has occurred, information on the location of the occurrence of the artifact, information on the type of the generated artifact, information on the degree of occurrence of the artifact, and the like.

According to an embodiment, the step of determining whether an artifact occurs in the medical image ( S3733 ) or the step of acquiring artifact information from the medical image ( S3735 ) may include performing using a learned neural network model. The step of determining whether an artifact has occurred in the medical image ( S3733 ) may include determining whether an artifact has occurred in the medical image using a learned neural network model. In the drawing, the step (S3733) of determining whether an artifact occurs in the medical image and the step (S3735) of acquiring artifact information from the medical image are expressed separately, but each step uses a single neural network model to determine whether or not an artifact occurs and information related to the artifact may include obtaining

The neural network model may acquire whether an artifact has occurred based on the medical image. The neural network model may acquire, based on the input medical image, whether an artifact exists in a specific region of the acquired medical image.

Also, the neural network model may acquire artifact information based on the medical image. Based on the input medical image, the neural network model may acquire information on whether or not an artifact has occurred, information about an artifact occurrence location, information about a type of the generated artifact, and information about an occurrence degree of an artifact.

The artifact information may be acquired based on a medical image acquired by the image acquisition device. Illustratively, the artifact information may be obtained using a neural network model trained to obtain the artifact information based on the medical image.

In order to obtain artifact information, a classifier algorithm that classifies or predicts an input image with respect to one or more labels may be used. For classification or prediction, various types of algorithms may be used. For example, k-nearest neighbor, support vector machine, artificial neural network, decision tree, self-organizing map, logical Logistic regression or the like may be used.

The artificial neural network may be a classifier, hybrid classifiers, ensemble classifiers, a linear regression neural network, or the like. The artificial neural network may include a Convolutional Neural Network (CNN). Artificial neural networks may be supervised learning, unsupervised learning, or reinforcement learning models.

25 is a diagram for describing a neural network model for acquiring artifact information according to an embodiment. Referring to FIG. 25 , a neural network model according to an embodiment may be provided in the form of a classifier. According to an embodiment, the neural network model includes an input layer (IL), a pooling layer (PL), a convolutional neural network layer (CL), a full-connection layer (FCL), a hidden layer (HL), an output layer (OL), and the like. , a feature vector may be obtained based on the input image. The neural network model may be prepared in the form of a classifier that classifies input images into one or more labels. Alternatively, the neural network model may be prepared in the form of a regression model. The neural network model may be prepared as a regression model that obtains a linear output value for specific artifact information based on an image as an input.

The neural network model may acquire output information by using a medical image obtained by photographing a specific region of the human body as an input. The output information may indicate whether the input image includes the target object. For example, the output layer of the neural network model may include an output node to which a probability function is assigned. The output layer of the neural network model may include an output node to which a probability function indicating whether the target image includes the target object is assigned. The output layer may include, for one or more target objects, an output node to which one or more probability functions indicating whether an input image includes each target object are assigned.

A neural network model can be trained to obtain artifact information. The neural network model may be trained to obtain artifact information based on training data including one or more medical images labeled with artifact information. The neural network model may be trained to obtain the artifact region based on training data including one or more medical images labeled with the artifact region.

The artifact learning data may include a plurality of medical images. The artifact learning data may include medical images taken in various ways, for example, CT, MRI, or X-ray images. The artifact training data may include a plurality of medical images including artifacts of various types, ranges, sizes, shapes, or locations.

The artifact learning data may include a medical image to which an artifact label indicating whether an artifact is generated or not. The artifact learning data may include a medical image obtained by photographing various parts of the human body and to which an artifact label indicating the occurrence of artifacts is attached. Here, the artifact label may be differently assigned according to the location of the occurrence of the artifact, the degree of occurrence of the artifact, the shape of the occurrence of the artifact, or the type of the artifact.

The artifact learning data may include a medical image masked in the artifact generation region. The artifact learning data may include masked (or labeled) medical images with respect to one or more artifact generating regions. The artifact learning data may include medical images that are masked (or labeled) differently with respect to a plurality of types of artifact generating regions. For example, the artifact learning data may include a medical image displayed in a first color for occurrence of a first type of artifact and displayed with a second color for occurrence of a second type of artifact.

The neural network model may be trained using the aforementioned artifact training data. The neural network model may be supervised, unsupervised, or reinforcement trained to obtain artifact information based on a medical image, using artifact learning data. A neural network model can be trained using a backpropagation method.

The neural network model may be trained to classify medical images according to whether or not they contain artifacts using artifact learning data including medical images labeled with or without artifacts. The neural network model may be trained to classify medical images according to types of artifacts included in the medical images. The neural network model can be trained to obtain whether a target image includes each type of artifact with respect to a plurality of artifacts through artifact learning data including medical images labeled with respect to the presence or absence of multiple types of artifacts.

The neural network model may be trained to acquire artifact region information through artifact learning data. The neural network model may be trained to detect an artifact-generating region from a target image by using artifact learning data including a medical image masked to the artifact-generating region with respect to one or more types of artifacts. The neural network model uses artifact learning data including a medical image labeled with respect to a plurality of types of artifact generation regions, and for each of a plurality of artifacts included in the medical image, a region and/or an artifact region for a type in which each artifact is distributed. can be learned to obtain information.

The neural network model may be trained using a plurality of artifact training data. The neural network model may be trained using first artifact training data including a medical image having a first type of artifact and second artifact training data including a medical image having a second type of artifact.

Meanwhile, a plurality of neural network models may be trained and used. The first neural network model may be trained based on the first artifact training data, and the second neural network model may be trained based on the second artifact training data. The second artifact training data may be at least partially different from the first artifact training data.

24 is a diagram for exemplarily explaining that the second image quality determination process determines the location of the artifact in the medical image using the artificial neural network model. Referring to FIG. 24 , the neural network model may acquire artifact region information regarding an artifact occurrence location by inputting a medical image acquired through an image acquisition device.

Referring to FIG. 24 , the step of acquiring artifact information from the medical image ( S3735 ) may include acquiring information about the artifact area in the medical image.

The artifact area information may be provided in the form of a mark for indicating a region where the artifact is likely to be located in the medical image, for example, in the form of a saliency map in the form of a heat map. The artifact area information may be a feature map obtained from a neural network model for determining whether or not artifacts are present or a neural network model for acquiring artifact information.

The artifact region may be a region in which relevance to the target artifact is greater than or equal to a reference value on the feature map based on a feature map obtained from the trained neural network model and related to the target artifact existing in the target medical image. The artifact area may include a first area and a second area positioned within the first area. The first region may be a region having a relevance to a target artifact equal to or greater than the first reference value on the feature map. The second area may be an area having a relevance to a target artifact on the feature map equal to or greater than a second reference value greater than the first reference value.

24 (a) to (c), the step of acquiring artifact information from the medical image (S3735) includes the first original image (a), the second original image (b), and the third original including artifacts. It may include acquiring an image (c). In this case, the first original image (a) to the third original image (c) may be images photographed on different planes, respectively. The first original image (a) to the third original image (c) may be images captured at the same point in time for the same subject.

24 (d) to (f), the step of obtaining artifact information from the medical image (S3735) is a first artifact image (d), a second artifact image (e), a second artifact image in which the position of the artifact is expressed 3 may include acquiring an artifact image (f). Here, the first artifact image (d) may be an image in which artifact area information obtained based on the first original image (a) is displayed in the first original image (a). The second artifact image (e) may be an image in which artifact area information obtained based on the second original image (b) is displayed in the second original image (b). The third artifact image f may be an image in which artifact area information obtained based on the third original image c is displayed in the third original image c.

Also, referring to (d) to (f) of FIG. 24 , the first to third artifact images are the first to fourth regions A1 to A4 in the order in which the artifacts are most likely to be located in the medical image. This may be the displayed image. Here, the first area A1 to the fourth area A4 may correspond to areas in which the artifact is located in the image. The first area A1 to the fourth area A4 may be an area including at least a portion of an area in which an artifact is located in the image. The first area A1 may be an area included in the second area A2 . The second area A2 may be an area included in the third area A3 . The third area A3 may be an area included in the fourth area A4 .

Although not shown in the drawing, the artifact area information may include a bounding box for displaying an area where the artifact is located in the medical image. The artifact area information may include coordinate information, pixel information, and the like at which the artifact is located in the medical image. Illustratively, the second image quality determining unit 3730 may use artificial neural network models such as Backpropagation Based Method (BBMs), Activation Based Method (ABMs), Pertubation Based Method (PBMs), etc. to obtain artifact occurrence location information. Here, the BBMs method may include LRP (Layer-wise Relevance Propagation), DeepLIFT, SmoothGrad, VarGrad, etc., and the ABMs method includes CAM (Class Activation Map), Grad-CAM (Gradient-Class Activation Map), etc. may exist, and as the PBMs method, there may be LIME (Local Interpretable Model-Agnostic Explanation), etc.

The third image quality determining unit 3750 according to an embodiment performs a quality determination in consideration of whether the segmentation information satisfies a quantitative criterion based on segmentation information obtained by anatomically segmenting a medical image obtained from the image acquisition device. can do. As a more specific example, the third image quality determining unit 3750 may determine whether at least some of the anatomical structures included in the medical image satisfy a predetermined quantitative criterion.

The third image quality determining unit 3750 may determine whether a value related to a predetermined region included in the image and corresponding to a part of the human body satisfies a criterion. The third image quality determining unit 3750 may determine whether at least a portion of the anatomical structure of the human body photographed through the image photographing apparatus is completely photographed.

For example, when a specific part of the human body is photographed by an image capturing device and then analyzed, the disease may be determined based on morphological indices of at least some of the photographed regions of the human body. If the imaging is not completely done, the result of the diagnosis of the disease cannot have a certain level of reliability. Therefore, since each region of the human body must be completely photographed, the third image quality determining unit 3750 considers whether the morphological index based on the anatomical region of the human body included in the medical image satisfies a generally required reference value. It may perform a function of determining whether the photographing of each region of the human body has been completely completed.

As a more specific example, when the image analysis apparatus 3000 analyzes an MRI image of the brain, the disease may be determined by calculating a ratio of volume values for each anatomical region of the brain. If the volume value of α does not satisfy the generally required standard value, the disease determination result cannot have a certain level of reliability or higher. In this case, if a specific area of the brain has not been completely photographed and the disease determination is made based on the volume value of the corresponding area, the result of the disease determination may be inaccurate. Accordingly, the third image quality determining unit 3750 determines whether the imaging of each region of the brain has been completely performed in consideration of whether the morphological index of each region of the brain included in the MRI image satisfies a generally required reference value. It can perform a function to determine whether or not

26 is a diagram for explaining a third image quality determination process.

Referring to FIG. 26 , the third image quality determination process includes an image segmentation information acquisition step (S3751), a quality determination target region acquisition step (S3753), a morphological index acquisition step of the target region (S3755), and a morphological index of the target region and comparing with a reference value (S3757).

The image segmentation information obtaining step S3751 may include obtaining segmentation information obtained based on the medical image segmented by the image segmentation module 3500 .

The segmentation information may include information about an anatomical structure of a human body obtained by segmenting a medical image. The segmentation information may include a morphological index of at least a partial region of an anatomical structure of a human body obtained by segmenting a medical image. Also, the segmentation information may include a morphological value of at least a partial region of an anatomical structure of a human body obtained by segmenting a medical image.

The morphological indicator may be obtained based on the morphological value. The morphological indicator may be obtained based on a plurality of morphological values, for example, a first morphological value and a second morphological value. For example, the morphological index may be obtained based on a ratio of the first morphological value to the second morphological value. In this case, the first morphological value may be the cerebellum, the second morphological value may be an anatomical region within the skull, and the morphological index obtained by the ratio of the first morphological value to the second morphological value is It may mean an ICV value. The morphological value may be information regarding an area, volume, location, or shape of at least a partial region of an anatomical structure of the human body.

The step of obtaining the quality determination target region ( S3753 ) may include specifying a target region on which a third image quality determination is to be performed based on the segmentation information. Here, the target region may include at least a partial region of an anatomical structure of a human body obtained through medical image segmentation. In addition, the target region may include at least a partial region that may affect the reliability of the disease determination result among the anatomical structures of the human body.

Exemplarily, the target region may include an outermost region from the center of the medical image among the anatomical structures of the human body acquired through medical image segmentation. The target region may include a region serving as a basis for disease diagnosis among anatomical structures of the human body acquired through medical image segmentation. The target region may include any one of the leftmost, right, upper, or lower regions of anatomical structures of the human body acquired through medical image segmentation. When the image quality is determined based on the medical image of the brain, the target region may include at least a part of an inner skull region of the brain obtained through medical image segmentation.

The step of obtaining the morphological index of the target region ( S3755 ) may include acquiring the morphological index of the specified target region. For example, the step of obtaining the morphological indicator of the target region ( S3755 ) may include acquiring the morphological indicator of the target region based on at least a part of a plurality of medical images obtained by the image acquisition device. The step of obtaining the morphological index of the target region ( S3755 ) may include acquiring a morphological value of the target region based on at least a portion of a plurality of medical images obtained by the image acquisition device.

Comparing the morphological index of the target region with a reference value ( S3757 ) may include comparing the morphological index or morphological value of the target region with a predetermined reference value. Here, the predetermined reference value may mean an average value of the morphological values of the target region for a plurality of ordinary people. The predetermined reference value may mean an index or a measurement value related to the target region when medical information having a certain level of reliability or higher can be obtained when diagnosing a disease based on the morphological indicator of the target region.

27 is a diagram for explaining an embodiment of a third image quality determination process.

Referring to FIGS. 27A to 27C , the third image quality determination process according to an exemplary embodiment may include performing quality determination based on a medical image of the brain. According to an embodiment, the third image quality determination process may include performing a quality determination as to whether the segmentation information satisfies a quantitative criterion based on segmentation information obtained by anatomically segmenting a medical image of the brain. have.

The image segmentation information acquisition step ( S3751 ) may include receiving segmentation information obtained by anatomically segmenting an MRI image of a brain. The quality determination target region acquisition step S3753 may include specifying a quality determination target region among each anatomically segmented brain regions, for example, a predetermined region corresponding to the cerebellum among each anatomically segmented brain regions. .

Exemplarily, when the quality determination target region is a predetermined region corresponding to the cerebellum, the segmentation information received in the image segmentation information acquisition step S3751 is based on an MRI image in which a predetermined region corresponding to a part of the brain is not completely photographed. It may be information obtained through Here, the criterion regarding whether the cerebellum has been completely photographed may be applied as the same criterion as the criterion for defining the cranial inner region based on the morphological value of the cerebellum in the part describing the ICV of the present specification.

Referring to (a) of FIG. 27 , in the image segmentation information acquisition step ( S3751 ), segmentation information obtained based on a first target region image in which at least a portion of a predetermined region corresponding to the cerebellum among each region of the brain is not photographed is obtained. may include receiving.

Referring to (b) of FIG. 27 , in the image segmentation information acquisition step (S3751), the segmentation information obtained based on the second target region image in which a predetermined region corresponding to the cerebellum among each region of the brain is all captured. may include

Referring to (c) of FIG. 27 , in the image segmentation information acquisition step ( S3751 ), all predetermined regions corresponding to the cerebellum among each region of the brain are photographed, but at least some of the regions corresponding to anatomical structures other than the cerebellum are completely photographed. and receiving segmentation information obtained based on the failed third target region image.

The step of obtaining the morphological index of the target region ( S3755 ) may include acquiring the morphological index of the quality determination target region, for example, a value of the volume of the cerebellum specified as the quality determination target region. The step of obtaining the morphological indicator of the target region ( S3755 ) may include comparing the morphological indicator of the target region with a reference value. The step of obtaining the morphological index of the target region ( S3755 ) may include determining whether a quantitative criterion is satisfied by comparing the volume value of the cerebellum specified as the target region with a predetermined reference value.

Exemplarily, the predetermined reference value may mean a morphological index of the cerebellum obtained based on an image of a second target region in which a predetermined region corresponding to the cerebellum is completely captured as shown in FIG. 27B . The predetermined reference value may be obtained based on a morphological index of the cerebellum obtained based on a plurality of images in which a predetermined region corresponding to the cerebellum is completely captured. The predetermined reference value may be an average value of cerebellar volume values obtained based on a plurality of images.

In this case, the step of comparing the morphological index of the target region with a reference value (S3757) is obtained based on the image of the first target region in which at least a portion of a predetermined region corresponding to the cerebellum is not photographed as shown in FIG. 27(a). Comparing the morphological index of the cerebellum with the morphological index of the cerebellum obtained based on the image of the second target region of FIG. can do. The step of comparing the morphological index of the target region with the reference value (S3757) determines whether the volume value of the cerebellum obtained based on the first target region image of FIG. and performing an image quality determination based on .

Referring to (c) of FIG. 27 , in the case of a third target region image in which the region corresponding to the cerebellum specified as the target region is completely photographed, but at least some of the regions corresponding to anatomical structures other than the cerebellum are not completely photographed, The third image quality determination process may include performing image quality determination after specifying the quality determination target region again.

Referring to (c) of Figure 27, when image quality is determined based on the region corresponding to the cerebellum, it is possible to determine the omission of the image inferior, but for the omission of the image superior It can be difficult to judge. In order to prevent such a problem, an index other than the region corresponding to the cerebellum may be used as an index for image quality determination. For example, the volume (or ICV) of a region corresponding to the skull or cerebrum may be used as an index of quality determination.

Alternatively, when a morphological index (eg, an ICV value of the cerebellum) obtained based on a region corresponding to the cerebellum is equal to or greater than a reference value, it may be determined that imaging of a region other than the cerebellum is incompletely performed.

The fourth image quality determining unit 3770 according to an embodiment may be configured to perform an image based on the relationship between segmentation information obtained by anatomically segmenting the medical image obtained from the image obtaining apparatus and artifact information extracted from the medical image obtained from the image obtaining apparatus. A quality judgment can be performed. For example, the fourth image quality determining unit 3770 may determine whether an artifact generated in the medical image overlaps with at least a part of anatomical structures included in the medical image.

For example, when the image analysis apparatus 3000 captures a specific part of the human body through the image photographing apparatus and then analyzes it, the disease may be determined based on segmentation information obtained by anatomically segmenting each area of the human body. , in this case, when an artifact is generated in segmentation information that can be a basis for disease determination, the result of disease determination based on the corresponding medical image may not have a certain level of reliability or higher. Conversely, even if an artifact is generated in a medical image, when the artifact does not affect segmentation information that can be a basis for disease determination, for example, when an artifact is generated in an area that does not overlap with the segmentation information, a disease based on the medical image The result of the judgment may have a reliability that satisfies a certain level. Therefore, the fourth image quality determining unit 3770 determines whether the medical image can be normally analyzed in consideration of whether the artifact affects the segmentation information that can be the basis for determining the disease even when an artifact is generated in the medical image. It can perform a function to determine whether or not

As a more specific example, when the image analysis apparatus 3000 analyzes an MRI image of the brain, the disease may be determined based on segmentation information obtained by anatomically segmenting each region of the brain. When an artifact is generated in at least a part of the segmented anatomical regions of the brain, the result of the disease determination based on the corresponding medical image may not have a certain level of reliability or higher. Accordingly, the fourth image quality determining unit 3770 determines whether the medical image can be normally analyzed in consideration of whether an artifact has occurred in at least a part of the anatomical structure of each region of the brain included in the MRI image. function can be performed.

28 is a diagram for explaining a fourth image quality determination process.

Referring to FIG. 28 , the fourth image quality determination process includes an artifact occurrence location information acquisition step (S3771), an image segmentation information acquisition step (S3773), a region of interest acquisition step (S3775), and an artifact occurrence location and interest It may include the step of determining the relationship of the region (S3777).

The step of acquiring artifact occurrence location information ( S3771 ) may include acquiring artifact occurrence location information. In this case, the artifact generation location information may be obtained from the image analysis apparatus 3000 . Artifact occurrence location information includes artifact area information on the location of artifact occurrence, for example, a heat map-type saliency map indicating an area where artifacts are likely to be located in a medical image, and displaying an area where artifacts are located in the medical image. may include a bounding box, coordinate information or pixel information on which an artifact is located in the medical image, and a detailed description thereof will be omitted since it has been described above in relation to the second image quality determination unit 3730 .

Artifacts referred to below may include noise, for example, various types of defects or conditions that affect acquisition of information based on an image. In addition, artifacts may include various image defects that are created as a result of improper image sampling in addition to image resolution and image brightness.

The image segmentation information obtaining step S3773 may include obtaining segmentation information obtained based on the medical image segmented by the image segmentation module 3500 . Here, the image segmentation information may include information obtained by anatomically segmenting a medical image. The image segmentation information may include segmentation information obtained by anatomically segmenting an MRI image obtained by photographing the brain, and a detailed description thereof will be omitted since it has been described above in relation to the third image quality determination unit 3750 . do.

The ROI-obtaining operation S3775 may include acquiring information related to the ROI based on segmentation information obtained by anatomically segmenting the medical image. The ROI-obtaining step S3775 may include acquiring information related to the ROI from among segmentation information obtained by anatomically segmenting a medical image. In the ROI-obtaining operation S3775, one or more regions may be obtained by segmenting the medical image, and a region corresponding to the ROI may be obtained.

The region of interest may include at least a partial region of an anatomical structure of a human body obtained through medical image segmentation. The region of interest may be predefined as a region corresponding to a specific anatomical (or functional) element on the medical image. The region of interest may include an outermost region from the center of the medical image among the anatomical structures of the human body acquired through medical image segmentation. The region of interest may include a region serving as a basis for disease diagnosis among anatomical structures of the human body acquired through medical image segmentation. The region of interest may include at least a portion of a cranial region and a cranial inner region of the brain obtained through medical image segmentation when image quality is determined based on an MRI image of the brain. The region of interest may be a factor related to diagnosis of a target disease.

The step of determining the relationship between the artifact occurrence position and the region of interest ( S3777 ) may include determining the relationship between the artifact occurrence position and the region of interest. In the step of determining the relationship between the location of the artifact and the region of interest ( S3777 ), quality determination is performed based on information on the region of interest among the anatomical structures of the human body obtained by segmenting the information on the artifact included in the medical image and the medical image. may include In the step of determining the relationship between the location of the artifact and the region of interest (S3777), the image quality may be determined based on whether the degree of overlap between the artifact generated in the medical image and the region of interest affects the reliability of the disease determination result below a certain level. have.

In the step of determining the relationship between the location of the artifact and the region of interest (S3777), it is determined whether at least some of the artifacts generated in the medical image are located on the region of interest, or whether at least some of the artifacts generated in the medical image overlap the region of interest. based on the quality judgment.

The step of determining the relationship between the artifact occurrence position and the region of interest ( S3777 ) may include determining the relationship between the artifact occurrence position and the region of interest based on the relationship between the region of interest and a baseline (artifact boundary) related to the position of the artifact. Here, the artifact boundary may be an outer line of the artifact area. An artifact boundary may be obtained from a neural network model that determines whether an artifact is present and based on a saliency map (eg, CAM) associated with estimating the location of the artifact.

In the step of determining the relationship between the location of the artifact and the region of interest ( S3777 ), an outer line of the artifact region and an outer line of the region of interest are obtained on the medical image, and whether the outer line of the artifact region and the outer line of the region of interest overlap on the medical image. based on , determining whether the artifact and the region of interest overlap. In this case, the image analysis apparatus 3000 may determine that the quality of the medical image is normal when there are less than two intersections between the outer line of the artifact area and the outer line of the ROI overlapping on the medical image. Alternatively, when there are two or more intersection points, the image analysis apparatus 3000 may determine that the quality of the target medical image is abnormal.

In the step of determining the relationship between the location of the artifact and the region of interest ( S3777 ), the number of common pixels included in both the artifact region and the region of interest is obtained from the medical image, and the overlapping degree of the artifact and the region of interest by using the number of common pixels may include judging In this case, the image analysis apparatus 3000 may determine that the quality of the medical image is abnormal when the number of common pixels exceeds a predetermined reference value. Alternatively, when the number of common pixels is less than or equal to a predetermined reference value, the image analysis apparatus 3000 may determine that the quality of the medical image is normal.

The step of determining the relationship between the location of the artifact and the region of interest ( S3777 ) may include determining whether an artifact has occurred in at least a portion of the region of interest, for example, the skull region or the skull inner region. The step of determining the relationship between the location of the artifact and the region of interest ( S3777 ) may include determining whether the artifact is overlapped in at least a portion of the region of interest, for example, the skull region and the skull inner region.

29 is a diagram illustrating an image in which an artifact has occurred. Hereinafter, an image quality determination process according to the occurrence range of the artifact will be described with reference to FIGS. 28 and 29 .

Referring to FIG. 29A , an image including artifacts overlapping the skull region and the internal region of the skull may be acquired. Referring to FIG. 29B , an image including an artifact overlapping a portion of the skull region may be acquired. Referring to FIGS. 29A and 29B , the region of interest may be an internal region of the skull or a region corresponding to the skull.

Referring to FIG. 29(a), when the target medical image includes the skull region and the artifact overlapping the internal region of the skull, and the region of interest is the internal region of the skull or the skull region, the relationship between the location of the artifact and the region of interest Determining ( S3777 ) may include determining that the region of interest and the artifact occurrence location overlap.

Referring to FIG. 29(b) , when the target medical image includes an artifact overlapping the skull region and the region of interest is the region inside the skull, determining the relationship between the location of the artifact and the region of interest (S3777) is the region of interest. and determining that the artifact occurrence location does not overlap.

Referring to FIG. 29(b) , when the target medical image includes an artifact that overlaps with the skull region, and the region of interest is the skull region, determining the relationship between the location of the artifact and the region of interest (S3777) includes the region of interest and the region of interest. It may include determining that the artifact occurrence positions overlap.

The artifact illustrated in FIG. 29 may mean an area determined as an artifact area. The region determined as the artifact region may be acquired through a segmentation neural network model acquiring the artifact region or a classifier model acquiring artifact information, and related content will be described in more detail below.

The fourth image quality determination may include acquiring a plurality of reference regions related to the artifact, and determining the quality of the image based on the overlap of each reference region and the ROI.

The step of determining the relationship between the artifact occurrence position and the region of interest ( S3777 ) according to an embodiment may include determining image quality based on artifact information related to the position of the artifact and the region in which the region of interest is distributed. The step of determining the relationship between the location of the artifact and the region of interest ( S3777 ) may include determining image quality based on information about a probability that an artifact is located in a specific region in the medical image and information on the region of interest. The step of determining the relationship between the location of the artifact and the region of interest ( S3777 ) may include determining whether an overlapping degree between the artifact generated in the medical image and the region of interest exceeds a predetermined reference value.

The information about the probability that the artifact is located in a specific region in the medical image may include artifact region information. Here, the artifact area information may include a first artifact area to an nth artifact area in an order of which artifacts are most likely to be located in the medical image. The artifact area may include a first area and a second area positioned within the first area as described above in the second image quality determination process, wherein the first area has a first reference value in relation to the target artifact on the feature map. It may be an area equal to or greater than the first reference value, and the second area may be an area having a relevance to a target artifact on the feature map greater than or equal to a second reference value.

30 is a diagram for explaining determining whether an artifact is generated to overlap an ROI in a fourth image quality determination process according to another exemplary embodiment;

Referring to (b) of FIG. 30 , the artifact area is a first artifact area (AR1), a second artifact area (AR2), a third artifact area (AR3), or A fourth artifact area AR4 may be included. The first artifact area AR1 to the fourth artifact area AR4 may include at least a portion of an area where the artifact is located in the image. The artifact regions may be the first artifact region AR1 , the second artifact region AR2 , the third artifact region AR3 , and the fourth artifact region AR4 in the order of showing a high correlation with the artifact.

Alternatively, the artifact area may include an area (eg, the third artifact area AR3 and the fourth artifact area AR4) having a relevance to the target artifact equal to or greater than the first reference value on the feature map. The artifact area may include an area (eg, the first artifact area AR1 and the second artifact area AR2) having a relevance to the target artifact on the feature map equal to or greater than a second reference value greater than the first reference value.

Referring to FIG. 30B , the overlapping degree of the artifact and the ROI may be determined using the number of common pixels included in both the artifact region and the ROI. For example, when the number of common pixels included in both the third artifact area A3 and the fourth artifact area A4 and the area of interest exceeds a predetermined reference value, the image analysis apparatus 3000 determines that the quality of the medical image is reduced. can be judged to be abnormal. As another example, when the number of common pixels included in both the first artifact area A1 and the second artifact area A2 and the area of interest is less than or equal to a predetermined reference value, the image analysis apparatus 3000 determines that the quality of the medical image is normal. can judge

Referring to FIG. 30C , the outer line of the artifact area may be used as an artifact boundary. An artifact boundary may be determined based on a feature map related to an artifact area as shown in FIG. 30A . The artifact boundary may be determined as a boundary of any one of the first artifact areas AR1 to AR4 as shown in FIG. 30B . An artifact boundary may be determined as an average of boundaries of the first artifact area AR1 to the fourth artifact area AR4 .

The artifact boundary may be an outer line of the first area (eg, an area having a relevance to a target artifact on the feature map that is equal to or greater than the first reference value). The artifact boundary may be an outer line of the second region (eg, a region having a relevance to a target artifact on the feature map that is greater than or equal to a second reference value greater than the first reference value).

Determining the relationship between the location of the artifact and the region of interest based on the relationship between the artifact boundary and the region of interest includes performing a quality judgment based on whether the artifact boundary and the region of interest overlap or not can do.

As shown in (c) of FIG. 30 , when there are two or more intersection points where the artifact boundary and the outer line of the ROI overlap on the medical image, the image analysis apparatus 3000 determines that the quality of the target medical image is abnormal. can

If there are a plurality of artifacts in the target image, the step of determining the relationship between the location of the artifact and the region of interest ( S3777 ) includes performing a quality determination based on whether any one of the plurality of artifacts and the region of interest overlap or the degree of overlap. can do.

The image output device may include an image quality related information output module 3900 . The image quality related information output module 3900 may output an image quality determination result performed by the image quality determination module 3700 of the image analysis apparatus 3000 . Also, the image quality related information output module 3900 may output information generated based on the image quality determination result performed by the image quality determination module 3700 of the image analysis apparatus 3000 .

The image output device or the image analysis device may obtain a user input instructing an additional operation in response to the selection window output.

31 is a diagram for describing an image quality related information output module according to an exemplary embodiment.

Referring to FIG. 31 , the image quality related information output module 3900 may include at least one of an error information output unit 3910 and a selection window output unit 3930 . In this case, the error information output unit 3910 may output error information regarding information obtained based on the image quality determination result. The selection window output unit 3930 may output a selection window regarding information obtained based on the image quality determination result.

The error information output unit 3910 and the selection window output unit 3930 may output information determined based on independent image quality determination results, respectively. For example, the error information output unit 3910 may output error information determined based on the first quality determination result, and the selection window output unit 3930 may output a selection window determined based on the second quality determination result.

Alternatively, the error information output unit 3910 and the selection window output unit 3930 may output information determined based on the same image quality determination result. The error information output unit 3910 may output error information determined based on the first quality determination result, and the selection window output unit 3930 may output a selection window determined based on the first quality determination result.

In FIG. 31 , the error information output unit 3910 and the selection window output unit 3930 are exemplified as being distinguished, but this is only an example, and the error information output unit 3910 and the selection window output unit 3930 are one unit. It may be provided in a physical or logical configuration. For example, the image quality related information output module 3900 may output error information generated based on the quality determination result and a selection window corresponding to the error information.

The selection window output unit 3930 may output a selection window for obtaining a user instruction related to an additionally performable operation based on the quality determination result.

The image quality related information output module 3900 may output at least one of error information and a selection window in text or image format.

The image quality related information output module 3900 may output error information. In this case, the error information may include information obtained based on the image quality determination result performed by the image quality determination module 3700 . In addition, the error information includes first error information related to the first image quality determination result, second error information related to the second image quality determination result, third error information related to the third image quality determination result, and the fourth image quality determination result It may include at least one of the fourth error information related to .

The image quality related information output module 3900 may output a first error information screen. The first error information may include information obtained based on the first image quality determination result. The first error information may include information obtained based on metadata information of the medical image.

According to an embodiment, the image quality-related information output module 3900 is configured to perform a first step when there is an error in the file structure information of the medical image obtained from the image acquisition device, for example, the file format of the medical image, the format of the medical image, or the size of the file. 1 Error information screen can be displayed. In this case, the first error information screen may include a first notification window indicating that there is an abnormality in the file structure information of the medical image or information about the file structure of the medical image.

According to another embodiment, the image quality-related information output module 3900 may include patient information on the medical image acquired from the image acquisition device, for example, information related to personal information such as the patient's name and age, and information related to the time the medical image was taken. When information or information about a patient's health state is missing, the first error information screen may be output. In this case, the first error information may include a first notification window indicating that patient information is missing from the medical image or patient information that is missing or included in the medical image.

32 is a diagram for exemplarily explaining a first error information screen.

Referring to (a) of FIG. 32 , the first error information screen is a medical image (MI) based on the first image quality determination, file structure information (DI) of the medical image, or patient information (PI) input to the medical image. may include at least one of

In this case, the medical image MI may be an image in which the medical image used to determine the first image quality is displayed. The file structure information DI of the medical image may include first file structure information regarding the format of the medical image, second file structure information regarding the format of the medical image, or third file structure information regarding the size of the medical image. . In this case, although not shown in the drawings, the first error information screen includes a detailed information object, and the output screen includes more detailed information about the medical image (MI) or file structure information (DI) according to the user's selection of the detailed information object. Content (eg, the patient's date of birth, the patient's address, the photographing time, the photographing location, the photographing device, etc.) may be additionally displayed.

Referring to FIG. 32B , the first error information screen may include a first notification window CO including first information indicating that the medical image is abnormal as a result of the first image quality determination. For example, the first notification window CO may include an error code indicating that there is an error in the file structure of the medical image or that patient information is missing from the medical image. In this case, the first error information screen includes a detailed information object, and the output screen includes detailed information about the first information according to a user selection for the detailed information object (eg, a description of an error code, generation of an error message) Causes, manuals that can respond to error messages, etc.) can be additionally displayed.

The image quality related information output module 3900 may output a second error information screen. The second error information may include information obtained based on the second image quality determination result. The second error information may include information obtained based on the noise information. The second error information may include information acquired based on various types of defects or conditions that affect acquisition of information based on the image.

According to an embodiment, the image quality related information output module 3900 may output the second error information screen when noise is included in the medical image acquired from the image acquisition device. The image quality related information output module 3900 may output the second error information in text or image format.

For example, the image quality related information output module 3900 may output the second error information screen when an artifact is included in the medical image acquired from the image acquisition device. The second error information screen may include an error message indicating that an artifact is included in the medical image. In addition, the second error information may include any one of information about a location where an artifact is generated in the medical image, information about an occurrence degree of the artifact, information about a range of occurrence of the artifact, or information about the type of the artifact.

The second error information screen may include text or an image. For example, the second error information screen may include a display, for example, an image in the form of a heat map, for indicating an area where the artifact is likely to be located in the medical image. As another example, the second error information screen may include an image indicating a bounding box indicating the location of the artifact in the medical image. As another example, the second error information screen may include an image for indicating anatomical segmentation information in the medical image.

33 and 34 are diagrams for exemplarily explaining the second error information screen.

Referring to FIG. 33A , the second error information screen may include a medical image (MI) display area related to the second image quality determination. The second error information screen may include an artifact information (AF) output area.

The medical image MI may be an image in which an area in which an artifact is likely to be located in the medical image is emphasized based on the second image quality determination. The medical image MI may be an image in which a region determined to be highly related to an artifact is emphasized in the form of a heat map. The medical image MI may be enlarged or reduced so that the type, occurrence range, shape, and the like of the artifact may be checked in more detail according to a user manipulation.

The artifact information AF may include at least one of information on the type of artifact, information on an area where the artifact is likely to be located, and information on the degree of occurrence of the artifact. The second error information screen includes a detailed information object, and the output screen includes more detailed information about the medical image (MI) or the artifact information (AF) according to the user selection for the detailed information object (eg, specific information in the image). information about the probability that the artifact is located in the region, heat map information about the occurrence range of the artifact, boundary information, etc.) may be additionally displayed.

Referring to FIG. 33B , the second error information screen may include a second notification window CO including second information on an artifact included in the medical image. For example, the second notification window CO may include a message in which an error code indicating that an artifact has occurred in the medical image is displayed. In this case, the second error information screen includes a detailed information object, and the output screen includes detailed information about the second information according to a user selection for the detailed information object (eg, a description of an error code, generation of an error message) Causes, manuals that can respond to error messages, etc.) can be additionally displayed.

Referring to FIG. 34 , the artifact information AF includes first artifact information AF1 including information on the type of artifact, and second artifact information AF2 including information on an area where the artifact is likely to be located. Alternatively, it may include the third artifact information AF3 including information on the degree of occurrence of the artifact.

The first artifact information AF1 may include information on the type of artifact, information on whether to correct the image in which the artifact is generated, and the like. In this case, the information on whether to correct the image in which the artifact is generated may include information obtained based on any one of the type of the artifact, the location of the occurrence, the range of occurrence, and the degree of occurrence of the artifact.

The second artifact information AF2 may include information on an area where an artifact is likely to be located in the medical image, and information on an analysisable area among regions corresponding to an anatomical structure in the medical image. In this case, the information on the analyzeable region may be obtained based on the anatomical segmentation information and the artifact occurrence location information. Illustratively, the information on the diagnosis region is information about a region that does not overlap an artifact among regions corresponding to the anatomical structure in the medical image or a region with a low degree of overlap with the artifact among regions corresponding to the anatomical structure in the medical image. information may be included. In this case, the diagnosable disease may vary according to the analysis possible region.

The third artifact information AF3 may include information about the occurrence degree, level, range, and the like of the artifact. In this case, the image analysis apparatus may determine whether to proceed with image analysis based on the third artifact information AF3 . For example, even when an artifact is generated in the medical image, the image analysis apparatus determines that the image analysis can be performed when it is determined that the degree of occurrence of the artifact is lower than a predetermined reference value based on the third artifact information AF3 can do.

The image quality related information output module 3900 may output a third error information screen. The third error information may include information obtained based on the third image quality determination result. Also, the third error information may include information obtained based on segmentation information obtained by anatomically segmenting a medical image obtained from the image obtaining apparatus.

According to an embodiment, the image quality-related information output module 3900 may display a third error information screen when at least a part of anatomical structures included in a medical image acquired from the image acquisition device does not satisfy a predetermined quantitative criterion. can be printed out. For example, the image quality related information output module 3900 may output a third error information screen when a value related to a predetermined region included in the medical image and corresponding to a part of the human body does not satisfy the criterion. As a more specific example, the image quality-related information output module 3900 may display a third error information screen when the volume value of a predetermined region corresponding to a part of the brain included in the MRI image of the brain does not satisfy a predetermined reference value. can be printed out. Also, the image quality related information output module 3900 may output the third error information screen in text or image format.

The third error information may include an error message indicating that there is an abnormality in at least a portion of anatomical structures in the medical image or information regarding an anatomical structure having an abnormality. As a more specific example, the third error information may include information related to a target region for which quality determination is to be performed based on the image segmentation result. The third error information may include information related to at least a partial region that may affect the reliability of the disease determination result among the anatomical structures of the human body acquired through image segmentation.

The third error information may include information about a morphological index or morphological value of at least a partial region of an anatomical structure of a human body obtained by segmenting a medical image, for example, an area, a volume, a location, or a shape.

The third error information may include information about a morphological index of the target region or a result of comparing a morphological value with a predetermined reference value. As a more specific example, the third error information may include information regarding whether at least a portion of anatomical structures included in the medical image satisfies a predetermined quantitative criterion.

The third error information screen may include text or an image. The third error information screen may include texts representing morphological indicators or morphological values of at least a portion of an anatomical structure of a human body obtained by segmenting a medical image. The third error information screen may include an image in which a target region corresponding to an anatomical structure that is a criterion for quality determination is displayed on the original medical image.

35 and 36 are diagrams for exemplarily explaining a third error information screen.

Referring to FIG. 35A , the third error information screen may include a medical image MI or region information RI related to a third image quality determination.

In this case, the medical image MI may include information related to the target region used for determining the third image quality based on the image segmentation result. For example, when it is determined whether the lower side of the medical image is completely photographed, the medical image may include an image in which information related to a target region for which quality determination is to be performed, for example, the cerebellum is expressed. The third error information screen may include a button for enlarging or reducing the medical image MI in order to check information on a region corresponding to an anatomical structure in the medical image in more detail. In addition, the region information (RI) includes any one of information related to a target region for which image quality determination is to be performed, information about a portion that is not completely captured in a medical image, or information on whether the target region meets a predetermined quantitative criterion. may include In this case, the third error information screen includes a detailed information object, and the output screen includes more detailed information about the medical image (MI) or area information (RI) according to the user selection for the detailed information object (eg, quality information on the region of the human body that has been the target of judgment, information on the image region that has been the target of quality judgment, etc.) may be additionally displayed.

Referring to FIG. 35B , the third error information screen may include a third notification window CO including third information on a region corresponding to an anatomical structure in a medical image. Illustratively, the third notification window CO may include a message in which error information regarding a region corresponding to an anatomical structure in the medical image, for example, an error code indicating that segmentation information does not satisfy a predetermined quantitative criterion is displayed. have. In this case, the third error information screen includes a detailed information object, and the output screen includes detailed information about the third information (eg, a description of an error code, generation of an error message according to a user selection for the detailed information object). Causes, manuals that can respond to error messages, etc.) can be additionally displayed.

Referring to FIG. 36 , the region information RI includes a first target region included in a first portion of a medical image and first region information ( RI1), the second target region included in the second part of the medical image, and the second region information RI2 including information on whether the second target region satisfies a predetermined quantitative criterion or an anatomical structure in the medical image It may include third area information RI3 including information on the arrangement of , and information on whether to correct or not.

For example, the first region information RI1 may include information on whether a morphological index or morphological value of a target region located above the medical image, for example, the frontal lobe satisfies a predetermined quantitative criterion. .

The second region information RI2 may include information on whether a morphological index or morphological value of a target region, eg, the cerebellum, located below the medical image satisfies a predetermined quantitative criterion.

The third region information RI3 includes information on the arrangement of anatomical structures in the medical image, for example, information on whether an anatomical structure in the medical image is photographed at an angle, and determination information on whether the medical image can be corrected based thereon. may include.

The image quality related information output module 3900 may output a fourth error information screen. The fourth error information may include information obtained based on the fourth image quality determination result.

According to an embodiment, the image quality related information output module 3900 may output a fourth error information screen when an artifact generated in the medical image overlaps with at least some of anatomical structures included in the medical image.

The fourth error information may include information obtained based on a relationship between segmentation information obtained by anatomically segmenting the medical image obtained from the image obtaining apparatus and artifact information extracted from the medical image obtained from the image obtaining apparatus. For example, the fourth error information may include information regarding whether an artifact has occurred in at least a portion of an anatomical structure of each region of the brain included in the MRI image.

The fourth error information may include information regarding whether an artifact generated in the medical image overlaps with at least a part of anatomical structures included in the medical image. In addition, the fourth error information may include information about an overlapping degree of at least a portion of an artifact generated in the medical image and an anatomical structure included in the medical image. In addition, the fourth error information may include information about the reliability of the disease determination result through analysis of the medical image based on the degree of overlap between the artifact generated in the medical image and at least a part of an anatomical structure included in the medical image. have.

The fourth error information may include information on a relationship between a probability that an artifact is located in a specific region in the medical image and a region of interest. In addition, the fourth error information includes a first artifact area, which is an area where artifacts are most likely to be located within the medical image, a second artifact area, which is an area where artifacts are most likely to be located within the medical image, or an artifact within the medical image. At least one of information about the third artifact area, which is an area that is unlikely to be located, may be included. The first to third artifact regions may be visually emphasized on the original medical image. The first to third artifact regions may be displayed in a heat map form to be superimposed on the original medical image.

The fourth error information screen may include text or an image. For example, the fourth error information screen may include artifact area information on an anatomical structure of a human body obtained by segmenting a medical image, for example, a saliency map in the form of a heat map indicating a region where an artifact is likely to be located in the medical image. , a bounding box for displaying a region in which an artifact is located in a medical image, and an image in which coordinate information or pixel information in which an artifact is located in the medical image is displayed.

37 and 38 are diagrams for exemplarily explaining a fourth error information screen.

Referring to (a) of FIG. 37 , the fourth error information screen may include at least one of a medical image (MI) related to the fourth image quality determination or information (QI) related to a relationship between an artifact and a region of interest. have.

The medical image MI may be an image in which a region of interest is displayed in the image. The medical image MI may be an image in which a region in which an artifact is highly likely to be located is displayed in a specific region within the image.

The medical image MI may indicate a degree of overlap between the ROI and the artifact acquired based on the location information of the artifact. For example, the medical image MI may display information related to the region where the ROI and the artifact overlap, for example, the area, shape, and number of pixels of the overlapped region. The medical image MI may indicate a ratio of the ROI and the region overlapping the artifact among the ROI.

The medical image MI may be an image in which a relationship between a region of interest and a probability that an artifact is located in a specific region in the medical image is displayed.

For example, the medical image MI may be an image in which an area corresponding to an artifact is visually emphasized on the target medical image. The medical image MI may be an image in which a region corresponding to an artifact is displayed on a target medical image in the form of a heat map, or a boundary line, eg, an outer line, of a region corresponding to the artifact is displayed.

Alternatively, the medical image MI may be an image displayed so that a portion where the artifact and the region of interest overlap is visually emphasized. The medical image MI may be an image in which a region corresponding to an artifact is displayed on a target medical image. The medical image MI may be an image in which a boundary line, for example, an outline, of a region corresponding to an artifact is displayed on the target medical image. The medical image MI may be an image in which a region corresponding to an artifact is displayed on a target medical image in a heat map format.

The information on the relationship between the artifact and the ROI may include, but is not limited to, information about the type of the ROI, overlapping information between the ROI and the artifact, or quality determination result information based on the overlapping information.

In this case, the fourth error information screen includes the detailed information object, and the output screen contains information about the medical image (MI) or information on the relationship between the artifact and the region of interest (QI) according to the user selection for the detailed information object. Detailed content (e.g., information expressing the degree of overlap between the region of interest and the artifact by dividing it into more granular steps, information indicating the degree of overlap between the region of interest and the artifact numerically, and details on the overlapping part with the artifact in the region of interest information, etc.) can be additionally displayed.

Referring to (b) of FIG. 37 , the fourth error information screen may include a fourth notification window CO including fourth information on the ROI in the medical image.

The fourth notification window CO may include a message in which an error code including information that an artifact overlaps with at least a portion of the ROI is displayed. In this case, the fourth error information screen includes a detailed information object, and the output screen includes detailed information about the fourth information (eg, a description of an error code, generation of an error message according to a user selection for the detailed information object). Causes, manuals that can respond to error messages, etc.) can be additionally displayed.

The fourth notification window CO includes an object for selecting whether to continue image analysis when at least a portion of the region of interest overlaps with an artifact, and the image analysis apparatus 3900 provides the user's response to the image analysis progress object. Depending on your choice, you can proceed with image analysis.

The fourth notification window CO includes an object that stops image analysis when the artifact overlaps with at least a portion of the region of interest, and the image analysis device 3900 stops image analysis according to the user's selection of the object to stop image analysis. can

Referring to FIG. 38 , information on the relationship between the artifact and the region of interest (QI) includes artifact information generated in a medical image (QI1), region of interest information (QI2) on a region corresponding to an anatomical structure in the medical image, or artifact. and at least one of information QI3 about a degree of overlap of a region corresponding to an anatomical structure. Illustratively, the fourth error information includes, when the region of interest is specified as a skull, information on whether an artifact is generated in a predetermined region corresponding to the skull, information on the degree of overlap of the artifact with a predetermined region corresponding to the skull, etc. may include

39 and 40 are diagrams for exemplarily explaining a selection window output by the image quality related information output module 3900 .

As described above, when inappropriate formal or substantial flaws exist in the acquired medical data, it may be difficult to obtain a reliability of a result of image analysis based on the medical data above a certain level.

However, even if a formal or substantial defect exists in the medical data, image analysis may be continuously performed when a predetermined criterion, for example, a criterion regarding the type or degree of defect present in the medical data is satisfied. Accordingly, even if there is a certain defect in the medical data, the image analysis is not uniformly stopped, but image analysis is decided according to a predetermined criterion, or the image analysis is performed after correcting or re-photographing the medical image. By allowing the process to be performed again, image analysis can be performed economically and efficiently in terms of time and cost.

The image quality determination process of the present invention may include, when a certain defect occurs in the medical data, providing a criterion for determining whether the image analysis apparatus can continue to analyze the image despite the defect. In addition, the image quality determination process provides information related to the quality of medical data to the user, and determines whether to proceed with image analysis according to the user's selection or determines whether to perform image analysis again after correcting or re-photographing the medical image. can provide a means to

Referring to FIG. 39 , the image quality related information output module 3900 may output a selection window based on the image quality determination result. The image quality related information output module 3900 may output a selection window when it is determined that the image quality is not normal through image quality determination. The image quality-related information output module 3900 may output image quality-related information after analyzing the image when it is determined that the image quality is normal through image quality determination.

The selection window may include a message generated based on the image quality determination. The selection window may include a message indicating that the quality of the medical image is not suitable for image analysis to be performed based on the image quality determination.

Referring to FIG. 40 , the image quality related information output module 3900 may output a selection window based on the image quality determination result. When it is determined that the image quality is not normal through image quality determination, the image quality related information output module 3900 may output a selection window for obtaining a user instruction related to image analysis progress. Accordingly, even when a defect occurs in the medical data, the image analysis apparatus may continue to analyze the image according to the user's selection based on a predetermined criterion. Even when a defect occurs in the medical data, the image analysis apparatus may correct or re-image the medical image according to a user's selection and then perform image analysis again.

The selection window may include a message generated based on the image quality determination result performed by the image quality determination module 3700 . The selection window may include a message generated based on results performed by the first image quality determining unit 3710 to the fourth image quality determining unit 3770 . The selection window may include a message generated based on the image quality determination. The selection window may include a message indicating that the quality of the medical image does not satisfy a certain level based on the image quality determination. The selection window may include a message indicating that the quality of the medical image is not suitable for image analysis to be performed based on the image quality determination.

The selection window may include a first button and a second button for obtaining whether the user agrees to the above-described message. The image analysis apparatus may output an additional message for selecting whether to proceed with image analysis according to the user's selection of the first button. The image analysis apparatus may output more detailed information about the message according to the user's selection of the second button.

The selection window may include a first selection window for obtaining a user instruction related to image analysis by the image analysis apparatus. That is, the image quality-related information output module 3900 performs the first selection for obtaining a user instruction as to whether or not to continue image analysis by the image analysis apparatus based on the result performed by the image quality determination module 3700 . window can be printed.

The selection window may include a second selection window for obtaining a user's instruction regarding the operation of re-performing the image quality determination process. That is, the image quality-related information output module 3900 outputs a second selection window for obtaining a user's instruction as to whether to proceed with the image quality determination process again based on the result performed by the image quality determination module 3700. can

According to an embodiment, when it is determined that an artifact is included in the medical image as a result of the quality determination by the second image quality determination unit 3730, the image quality related information output module 3900 may output a first selection window. have.

The first selection window may output a first message regarding artifact information included in the medical image, for example, whether or not an artifact has occurred, an artifact occurrence location, an artifact occurrence degree, an artifact occurrence range, or an artifact type.

When it is determined that an artifact is included in the medical image as a result of the quality determination by the second image quality determination unit 3730, the image quality related information output module 3900 obtains a user's instruction related to re-performing the image quality determination process A second selection window can be output.

The second selection window may output a second message requesting a user's instruction as to whether to proceed with the image quality determination process again after correcting artifacts included in the medical image. The second selection window may output a second message requesting a user's instruction as to whether or not to proceed with the image quality determination process again after re-photographing the medical image.

According to another embodiment, when it is determined that the segmentation information does not satisfy a quantitative criterion based on the segmentation information obtained by anatomically segmenting a medical image as a result of quality determination by the third image quality determination unit 3750, image quality related The information output module 3900 may output a first selection window.

The first selection window may include a morphological index or morphological value of at least a partial region among regions obtained through medical image segmentation, for example, a first message regarding an area, volume, location, or shape of the corresponding region.

When it is determined that the segmentation information does not satisfy the quantitative criterion based on the segmentation information obtained by anatomically segmenting the medical image as a result of the quality determination by the third image quality determination unit 3750, the image quality related information output module 3900 may output a second selection window for obtaining a user's instruction related to a re-performing operation of the image quality determination process.

The second selection window may output a second message requesting a user instruction regarding whether to proceed with the image quality determination process again after correcting the anatomical segmentation information in the medical image. The second selection window may output a second message requesting a user's instruction as to whether or not to proceed with the image quality determination process again after re-photographing the medical image.

According to an embodiment, when an artifact generated in the medical image is overlapped with at least a part of anatomical structures included in the medical image as a result of quality determination by the fourth image quality determining unit 3770 , the image quality related information output module 3900 may output a first selection window.

The first selection window may output a first message regarding relationship information between the artifact and the region of interest included in the medical image, for example, the degree of overlap between the artifact and the region of interest, location information of the artifact existing in the region of interest, and the like.

As a result of the quality determination by the fourth image quality determination unit 3770 , when an artifact generated in the medical image overlaps with at least a portion of anatomical structures included in the medical image, the image quality related information output module 3900 provides a user A second selection window for obtaining a user's instruction related to re-performing the image quality determination process may be output.

The second selection window may output a second message requesting a user's instruction regarding whether to proceed with the image quality determination process again after correcting at least a part of anatomical structures included in the medical image. The second selection window may output a second message requesting a user's instruction as to whether or not to proceed with the image quality determination process again after re-photographing the medical image.

Hereinafter, a quality determination process according to various embodiments will be described. Unless otherwise specified, the contents of each step may be similarly applied to the contents described above in this specification.

41 to 43 are diagrams for explaining first to third quality determination processes. In the first to fourth quality determination processes, data as a standard for quality determination, a quality standard as a standard for quality determination, and output quality information may be different.

41 to 43 , the first to third quality determination processes include a medical image acquisition step (S3000), a medical image quality determination step (S3300), a medical image preprocessing step (S3200), and a medical image segmentation step (S3400). Alternatively, the step of outputting medical image quality-related information ( S3500 ) may be included.

Referring to FIG. 41 , the first quality determination process includes a medical image acquisition step (S3000), a medical image quality determination step (S3300), a medical image preprocessing step (S3200), a medical image segmentation step (S3400), and medical image quality related information It may be performed in the order of the output step (S3500). In this case, the medical image quality determination step S3300 may include determining the image quality based on the raw data before the pre-processing operation. The medical image preprocessing step S3200 may include performing preprocessing on the image on which the medical image quality determination has been performed.

42 , the second quality determination process includes a medical image acquisition step (S3000), a medical image preprocessing step (S3200), a medical image quality determination step (S3300), a medical image segmentation step (S3400), and medical image quality related information It may be performed in the order of the output step (S3500). In this case, the medical image quality determination step S3300 may include determining the image quality based on the pre-processed image. The medical image segmentation step S3400 may include performing segmentation on the image on which the medical image quality determination has been performed.

Referring to FIG. 43 , the third quality determination process includes a medical image acquisition step (S3000), a medical image preprocessing step (S3200), a medical image segmentation step (S3400), a medical image quality determination step (S3300), and medical image quality related information It may be performed in the order of the output step (S3500). In this case, the medical image quality determination step S3300 may include determining the image quality based on the image on which the pre-processing operation and the segmentation operation have been performed.

44 and 45 are diagrams for explaining a fourth quality determination process.

Referring to FIG. 44 , the fourth quality determination process includes a medical image acquisition step (S3000), a first image quality determination step (S3310), a medical image preprocessing step (S3200), a medical image segmentation step (S3400), and a second image quality It may include a determination step (S3320), a third image quality determination step (S3330), a fourth image quality determination step (S3340), and a medical image quality related information output step (S3500).

The first image quality determination step S3310 may include performing a first image quality determination based on raw data before the pre-processing operation. The second to fourth image quality determination steps may include performing a second image quality determination based on an image on which a preprocessing operation and a segmentation operation have been performed. In this case, the order of the second to fourth image quality determination steps does not limit the invention described in this specification, and the order of each step may be changed.

For example, after the third image quality determination step S3330 is performed based on the image on which the preprocessing operation and the segmentation operation have been performed, the second image quality determination step S3320 may be performed. Alternatively, the second image quality determination step S3320 and the third image quality determination step S3330 may be performed in parallel based on the image on which the preprocessing operation and the segmentation operation have been performed.

Referring to FIG. 45 , the image analysis apparatus 3000 may acquire a medical image from the image acquisition apparatus ( S3910 ). The image analysis apparatus 3000 may perform a primary image quality determination based on the acquired medical image ( S3911 ). Here, the primary image quality determination may include a series of quality determinations performed before the medical image segmentation operation. For example, the primary image quality determination may include a quality determination performed based on metadata of a medical image.

In this case, when it is determined that the image quality does not satisfy the predetermined condition as a result of the primary image quality determination, the image analysis apparatus 3000 may stop the image analysis and output medical image quality related information ( S3918 ).

When it is determined that the image quality satisfies a predetermined condition as a result of the primary image quality determination, the image analysis apparatus 3000 may perform medical image segmentation ( S3913 ). Also, the image analysis apparatus 3000 may determine the secondary image quality based on the medical image on which the segmentation operation is performed ( S3914 ).

As a result of the secondary image quality determination, if it is determined that no artifact is generated in the ROI in the medical image, the image analysis apparatus 3000 continues to analyze the medical image (S3917) and then outputs information related to the medical image quality (S3918) can do.

When it is determined that the artifact is generated in the ROI in the medical image as a result of the secondary image quality determination, the image analysis apparatus 3000 may stop the image analysis. In this case, the image analysis apparatus 3000 may output medical image quality related information after stopping the image analysis ( S3918 ). For example, when it is determined that an artifact is generated in the region of interest in the medical image as a result of the secondary image quality determination, the image analysis apparatus 3000 stops the image analysis and corrects the medical image and then performs the image quality determination process. can be done again. As another example, the image analysis apparatus 3000 may retake a medical image after stopping the image analysis and then perform the image quality determination process again.

If it is determined as a result of the secondary image quality determination that no artifact is generated in the region of interest in the medical image, the image analysis apparatus 3000 continues to perform image analysis (S3917) and then outputs medical image quality related information (S3918). can

46 and 47 are diagrams for explaining a fifth quality determination process.

Referring to FIG. 46 , the fifth quality determination process includes a medical image acquisition step (S3000), a first image quality determination step (S3310), a medical image preprocessing step (S3200), a second image quality determination step (S3320), and a medical image It may include a segmentation step (S3400), a third image quality determination step (S3330), a fourth image quality determination step (S3340), and a medical image quality related information output step (S3500).

The first image quality determination step S3310 may include performing a first image quality determination based on raw data before the pre-processing operation. The second image quality determination step S3320 may include quality determination based on the image on which the pre-processing operation has been performed. The third image quality determination step S3330 and the fourth image quality determination step S3340 may include quality determination based on the image on which the segmentation operation has been performed.

The order of the third image quality determination step S3330 and the fourth image quality determination step S3340 does not limit the invention described herein, and the order of each step may be changed. For example, the third image quality determination step S3330 and the fourth image quality determination step S3340 may be performed in parallel based on the image on which the preprocessing operation and the segmentation operation have been performed.

Referring to FIG. 47 , the image analysis apparatus 3000 may acquire a medical image from the image acquisition apparatus ( S3920 ). The image analysis apparatus 3000 may determine the primary image quality based on the acquired medical image ( S3921 ).

When it is determined that the image quality does not satisfy the predetermined condition as a result of the primary image quality determination, the image analysis apparatus 3000 may stop the image analysis and output medical image quality related information ( S3929 ).

When it is determined that the image quality satisfies a predetermined condition as a result of the primary image quality determination, the image analysis apparatus 3000 may perform a secondary image quality determination ( S3923 ). Here, the secondary image quality determination may include a series of quality determinations performed before the image segmentation operation.

The image analysis apparatus 3000 may perform medical image segmentation based on the medical image on which the secondary image quality determination is performed ( S3924 ). Also, the image analysis apparatus 3000 may perform a tertiary image quality determination based on segmentation information on which medical image segmentation has been performed ( 3925 ). Here, the tertiary image quality determination may include a series of quality determinations performed after the image segmentation operation.

When it is determined that no artifact is generated in the ROI in the medical image as a result of the tertiary image quality determination, the image analysis apparatus 3000 may continue to analyze the medical image and then output medical image quality related information ( S3929 ). .

When it is determined that an artifact is generated in the ROI in the medical image as a result of the third image quality determination, the image analysis apparatus 3000 may stop the image analysis 3000 and output medical image quality related information ( S3929 ). Here, the medical image quality-related information may include information based on primary to tertiary image quality determination.

If it is determined as a result of the tertiary image quality that no artifact is generated in the ROI in the medical image, the image analysis apparatus 3000 may continue to analyze the image and then output medical image quality related information ( S3929 ). Here, the medical image quality-related information may include information based on primary to tertiary image quality determination.

The image quality determination process according to the present specification may be implemented in an order other than the order described in FIGS. 44 to 47 .

With respect to brain-related diseases, particularly dementia, the degree of brain contraction is indexed and used as an auxiliary index for diagnosing dementia. There are various indicators related to the degree of brain contraction, and an intracranial volume analysis (ICV Analysis) is used to provide an auxiliary indicator related to the diagnosis of dementia as an example to obtain the indicators. However, in the conventional analysis of the internal volume of the skull, a method of matching the target image of the subject to a standard brain model was used. However, there was a problem that the method of matching the target image of the object to the standard brain model could be inaccurate because the volume analysis considering race was not properly reflected due to differences in the standard brain model between races. In addition, as the individual brain structure is matched to the standard brain model, there is a problem that there is a limit to the volume analysis that reflects the characteristics of the individual brain.

The image analysis apparatus 2000 according to an embodiment of the present application may perform segmentation of the target image without matching the target image of the object to a standard brain model to acquire the skull and its internal region of the object, based on this It can provide the advantageous effect that the accuracy of the analysis of the cranial internal volume can be improved.

Hereinafter, a volume analysis method implemented by the image analysis apparatus 2000 according to an embodiment of the present application will be described in detail with reference to FIGS. 48 to 57 .

See FIG. 48 . 48 is a flowchart illustrating an operation of an image analysis method implemented by the image analysis apparatus 2000 according to an embodiment of the present application. Specifically, FIG. 48 is a flowchart illustrating an operation of acquiring a morphological index of a brain implemented by the image analysis apparatus 2000 according to an embodiment of the present application.

An image analysis method according to an embodiment of the present application includes: acquiring an image (S4100); It may include a pre-processing step of the image, an image alignment step (S4200), an image segmentation step (S4300), and a step (S4400) of obtaining a morphological index.

In the step of acquiring an image ( S4100 ), the image analysis apparatus 2000 may acquire a target image from the image acquiring apparatus 1000 .

In the image pre-processing and image alignment step ( S4200 ), the above-described image pre-processing operations may be performed. For example, in the image pre-processing and image alignment step ( S4200 ), the image analysis apparatus 2000 pre-processes the target image, such as converting the format of the target image, removing noise from the image, or correcting the intensity of the image. It can be implemented to perform an action.

In addition, in the image pre-processing and image alignment step S4200, the image alignment operations described above with reference to FIGS. 5 to 6 may be performed. By aligning the target image, the image analysis apparatus 2000 according to an embodiment of the present application may input a target image having a direction corresponding to the direction of the image of the training data of the neural network model to the neural network model, and thus a unified standard Analysis results can be derived under

As an example, the image analysis apparatus 2000 may be implemented to align the orientation of the image based on data related to the orientation of the brain image structured as metadata for the target image. For example, the image analysis apparatus 2000 matches a photographed brain image with respect to left and right, anterior and posterior, and superior and inferior, respectively, of the object by using metadata of the target image. image alignment operation can be performed.

As another example, the image analysis apparatus 2000 may be implemented to perform spatial normalization of a target image as a method of an image alignment operation. Specifically, the image analysis apparatus 2000 may be implemented to perform spatial normalization of the target image based on the brain template. More specifically, the image analysis apparatus 2000 may convert the coordinates of the target image so that the spatial distribution of the target image becomes optimal for the artificial neural network model by matching the target image to the brain template, and align the target image.

As an example, the image analysis apparatus 2000 may be implemented to align the target image based on the feature region in the target image.

See FIG. 49 . 49 is an exemplary diagram of an image alignment method implemented by the image analysis apparatus 2000 according to an embodiment of the present application.

For example, referring to FIG. 49 , feature regions corresponding to an anterior commissure (AC) and a posterior commissure (PC) may exist in the brain image. In this case, the image analysis apparatus 2000 may be implemented to perform spatial alignment of the target image based on the AC-PC line connecting the anterior commissure AC and the posterior commissure PC.

For example, the first image may be photographed in the first direction, the second image may be photographed in the second direction, and the third image may be photographed in the third direction. Since the images were taken in a random direction, the images acquired from the image acquisition apparatus 1000 may not be spatially aligned. In this case, the image analysis apparatus 2000 may unify the directions of the images by spatially aligning the images.

For example, the first image may be acquired with the AC-PC line L1 connecting the anterior commissure AC1 and the posterior commissure PC1 facing in the first direction, and the second image may be obtained with the anterior commissure AC2 and An AC-PC line (L2) connecting the posterior commissure (PC2) may be acquired in a state in which it is directed in the second direction, and the third image is an AC-PC line connecting the anterior commissure (AC3) and the posterior commissure (PC3). (L3) may be obtained in a state facing in the third direction.

In this case, the image analysis apparatus 2000 may be implemented to align the images so that the AC-PC lines L1 , L2 , and L3 of each image spatially face the same reference direction.

For example, the image analysis apparatus 2000 may align the second image and the third image by using the AC-PC line L1 from which the first image is obtained as a reference direction.

As another example, the image analysis apparatus 2000 may obtain an input for a reference direction from the user, and may align the first image, the second image, and the third image based on the user's input.

However, the above description is only an example, and the target image may be aligned with respect to an arbitrary reference direction so that the image analysis apparatus 2000 unifies the direction of the target image to be analyzed.

In addition, as described above, although the feature region in the image has been described as anterior commissure and posterior commissure, this is only an example, and any suitable feature region for performing spatial alignment of the target image may be utilized. For example, the image analysis apparatus 2000 may be implemented to align the images so that the center of the two eyes and the line connecting the centers of the two eyes face the same direction.

In addition, the image analysis apparatus 2000 according to an embodiment of the present application may be implemented to align the target image by matching the target image with respect to the template generated based on the above-described AC-PC line (or plane). . In other words, although the method has been described in which the target image is aligned so that the AC-PC lines of the images included in the target image are in the same direction with reference to FIG. may be created, and may also align the target image by matching the target image to a standard template.

The image analysis apparatus 2000 according to an embodiment of the present application may improve segmentation accuracy of a target image using an artificial neural network by performing spatial alignment of the target image.

In addition, in the case of volume analysis such as intracranial volume analysis (ICV Analysis), it is necessary to define an internal region for calculating a volume value. Since it can be implemented to define an inner region for an image, the reproducibility or accuracy of volume analysis can be improved.

In the image segmentation step S4300 , the image analysis apparatus 2000 may segment the target image according to the image segmentation operation described above with reference to FIGS. 7 to 17 .

을 획득하도록 구현될 수 있다. 이때, 형태학적 지표를 계산하기 위하여는, 이미지 분석 장치(2000)는 대상 이미지로부터 적어도 타겟 요소에 대응되는 영역과 제2 내부 영역에 대응되는 영역을 획득하여야 한다.The image analysis apparatus 2000 according to an embodiment of the present application is an example of a morphological index to be described later,

can be implemented to obtain In this case, in order to calculate the morphological index, the image analysis apparatus 2000 must obtain at least a region corresponding to the target element and a region corresponding to the second inner region from the target image.

In this case, the second internal region may be a region included in the internal region of the skull. Accordingly, the image analysis apparatus 2000 may be implemented to acquire at least a region corresponding to the skull and an internal region of the skull from the target image.

Also, the image analysis apparatus 2000 may acquire the second internal region by correcting a boundary related to the internal region of the skull obtained from the target image. At this time, in order to correct the boundary related to the inner region of the skull, the reference region may be the basis. Accordingly, the image analysis apparatus 2000 may be implemented to obtain a region corresponding to the reference region from the target image.

Accordingly, the target image may be divided into a plurality of regions including at least a skull region and a region corresponding to the target element by the segmentation operation of the image analysis apparatus 2000 . Also, a reference region serving as a reference for correcting a boundary of an internal region of the skull, which will be described later, may be additionally acquired by the segmentation operation of the image analysis apparatus 2000 .

For example, the reference region as a reference for correcting the boundary of the inner region of the skull may be a region corresponding to the cerebellum or a region corresponding to the cervical vertebrae. However, this is only an example, and it is obvious that a region corresponding to an arbitrary brain element may be acquired as a reference region as needed to define the internal region of the skull.

The image analysis apparatus 2000 according to an embodiment of the present application may acquire the inner region of the skull based on the region of the skull obtained by segmentation.

According to the image analysis method according to an embodiment of the present application, the image analysis apparatus 2000 may calculate a morphological index ( S4400 ). In this case, the morphological index may be interpreted as information including quantitative information or qualitative information that may be obtained from an image.

For example, according to the image analysis method according to an embodiment of the present application, a morphological index related to a brain volume may be obtained as a morphological index.

The shape of the morphological indicators associated with the volume of the brain may vary.

For example, the morphological index may include a normalized brain parenchymal volume (NBV), a regional brain parenchymal volume (RBPV), and the like.

으로 정의될 수 있으며, ICV와 관련된 형태학적 지표는 뇌질환(예, 우울증, 치매) 진단의 보조 지표로 사용될 수 있다. For another example, according to the image analysis method according to an embodiment of the present application, it may be implemented to calculate a morphological index related to ICV. At this time, the morphological indicators related to ICV are

It can be defined as , and morphological indicators related to ICV can be used as auxiliary indicators for diagnosing brain diseases (eg, depression, dementia).

Hereinafter, an image analysis method according to an embodiment of the present application for calculating a morphological index related to improved accuracy and personalized ICV will be mainly described.

In the step of calculating the morphological index ( S4400 ), the image analysis apparatus 2000 may be implemented to correct the boundary of the inner region of the skull and to calculate the morphological index based on the corrected boundary. .

Hereinafter, the step of calculating the morphological index ( S4400 ) will be described in more detail with reference to FIG. 50 . FIG. 50 is a diagram illustrating a flowchart of an image analysis method according to an embodiment of the present application, and in detail, the step S4400 of FIG. 48 is subdivided.

Referring to Figure 50, the step of calculating the morphological index according to an embodiment of the present application (S4400), the step of obtaining the first boundary of the first internal region of the skull (S4410), by modifying the first boundary Obtaining a second internal region having a second boundary (S4420), acquiring a morphological volume value of the second internal region (S4430), acquiring a morphological volume value of a region corresponding to a target element (S4430) S4440) and calculating the morphological volume index (S4450) may be further included.

Specifically, in the step of calculating the morphological index (S4400), the image analysis apparatus 2000 is based on the skull area obtained in the image segmentation step (S4300) in the first internal region of the skull and the first internal region of the skull. A corresponding first boundary may be acquired.

See FIG. 51 . 51 is a diagram illustrating an example of a method of correcting the first boundary of the first inner region of the skull according to an embodiment of the present application.

Referring to FIG. 51 , the image analysis apparatus 2000 may acquire the first internal region IR1 of the skull based on the skull region SR obtained in the image segmentation step S4300 . Also, the image analysis apparatus 2000 may acquire a first boundary BD1 corresponding to the first internal region IR1 of the skull based on the acquired first internal region IR1 of the skull.

As an example, the image analysis apparatus 2000 may acquire a region corresponding to the cerebrospinal fluid (CSF) in the image segmentation step S4300, and may acquire a skull region based on the region corresponding to the cerebrospinal fluid (CSF). have. For example, the image analysis apparatus 2000 may acquire the skull region by using the outer boundary of the region corresponding to the cerebrospinal fluid (CSF) as the inner boundary of the skull region.

In this case, the image analysis apparatus 2000 may acquire the outer boundary of the region corresponding to the cerebrospinal fluid CSF as the first boundary BD1 .

Alternatively, the image analysis apparatus 2000 may acquire an inner boundary of a region corresponding to the cerebrospinal fluid CSF as the first boundary BD1 .

However, the above-described first boundary is merely an example, and the image analysis apparatus 2000 is a first boundary (BD1) composed of arbitrary regions between the external boundary and the internal boundary of the region corresponding to the cerebrospinal fluid (CSF). can be obtained

The image analysis apparatus 2000 according to an embodiment of the present application may smooth or correct a boundary corresponding to the obtained internal region of the skull (eg, a first internal region or a second internal region to be described later). There may be a possibility that the boundaries corresponding to the specifically obtained inner region of the skull are not clear or there is an error. Accordingly, the image analysis apparatus 2000 may smooth a boundary corresponding to a boundary corresponding to the obtained internal region of the skull or correct an error related to the boundary.

For example, when the boundary is not clear, the image analysis apparatus 2000 may be implemented to use any suitable image intensity correction method and smoothing method, such as the above-described image preprocessing method.

For example, when an error exists in the obtained boundary, the image analysis apparatus 2000 may obtain a user input for correcting the boundary through the input module, and may be implemented to correct the boundary based on the user input. Alternatively, the image analysis apparatus 2000 may be implemented to use any suitable noise removal method, such as the image pre-processing technique described above.

The image analysis apparatus 2000 according to an embodiment of the present application provides a first boundary BD1 related to the first internal region IR1 of the skull based on the reference region RR obtained in the image segmentation step S4300. An operation to correct some may be performed.

For each target image, the location or size of a region corresponding to the brain included in the acquired image may be different according to a photographing method of the image.

For example, a brain region included in the first internal region of the skull acquired by the first imaging method may be different from a brain region included in the first internal region of the skull acquired by the second imaging scheme.

Therefore, there is a need for a boundary definition method that includes a region necessary to provide information related to disease diagnosis assistance and is not affected by an image capturing method.

The image analysis apparatus 2000 according to an embodiment of the present application may correct a part of the first boundary corresponding to the first internal region of the skull, and based on the “common standard” for the target image, the first boundary Since an operation for correcting a part can be performed, meaningful information for disease diagnosis assistance can be acquired, and disease diagnosis assistance indicators can be acquired according to a common criterion for each target image.

Hereinafter, an example of a method of correcting the first boundary of the first inner region of the skull according to an embodiment of the present application will be described with reference to FIGS. 51 and 52 . 52 is a diagram illustrating an example of a second internal region obtained according to a method of correcting a first boundary of a first internal region of a skull according to an embodiment of the present application.

The first boundary BD1 obtained in the above-described image segmentation step S4300 may be a 3D boundary.

In this case, the image analysis apparatus 2000 corrects a 3D first boundary in order to obtain a region necessary for providing meaningful information to aid in disease diagnosis, and corrects a reference plane (or 2D image slice, RP) can be obtained. Also, in order to obtain the reference plane RP, the image analysis apparatus 2000 may use a feature region included in a sagittal plane image as shown in FIGS. 51 and 52 .

For example, the image analysis apparatus 2000 may acquire the reference plane RP related to the reference region RR. The image analysis apparatus 2000 may acquire the reference plane RP adjacent to the reference region RR from the target image. For example, the image analysis apparatus 2000 may obtain a plane adjacent to a lower edge of a boundary defining the reference region RR as the reference plane RP. The reference plane RP may be a plane parallel to a transverse plane of the target image and perpendicular to a sagittal plane. In this case, the reference region RR may be a region corresponding to the cerebellum.

The image analysis apparatus 2000 may obtain the second boundary BD2 by correcting a part of the first boundary of the first inner region based on the obtained reference plane RP. For example, the image analysis apparatus 2000 corrects a portion of the first boundary BD1 positioned on the lower side of the reference plane RP and the first positioned on the upper side of the reference plane RP. The second boundary BD2 may be acquired based on the boundary BD1 .

More specifically, the image analysis apparatus 2000 is implemented to correct the first boundary by replacing a portion of the first boundary located below the reference plane RP with a reference line or reference plane included in the reference plane RP. can be

The image analysis apparatus 2000 may acquire the second boundary by correcting the first boundary, and may acquire the second internal region of the skull having the second boundary.

The image analysis apparatus 2000 according to an exemplary embodiment of the present application modifies a portion of the first boundary BD1 based on the reference plane RP, thereby providing a second region necessary for providing information related to disease diagnosis assistance. Since the internal region BD2 can be acquired and the morphological index can be calculated based on the second internal region BD2 obtained according to a common criterion for the target image, meaningful information can be obtained to assist in disease diagnosis. have.

Referring to FIG. 52 , the image analysis apparatus 2000 is configured based on a reference plane (RP, for example, a reference plane including a lower edge of the reference region RR) adjacent to the reference region RR. By modifying a part of the first boundary BD1 , the second inner region IR2 of the skull having the second boundary BD2 may be obtained. At this time, compared with FIG. 51 , it can be confirmed that the portion of the first boundary BD1 below the reference plane RP is replaced with the reference plane of the reference plane RP, so that the first boundary BD1 is corrected. . Also, by modifying the first boundary BD1 , the second inner region IR2 having the second boundary BD may be obtained.

51 and 52 show an example of correcting a part of the first boundary BD1 of the first internal region IR of the skull only for a brain image related to a sagittal plane among brain images, but the image analysis apparatus (2000) may also perform a corresponding boundary correction operation on a brain image related to a coronal plane and a transverse plane among the brain images.

Hereinafter, another example of a method of correcting the first boundary of the first inner region of the skull according to an embodiment of the present application will be described with reference to FIG. 53 . 53 is a diagram illustrating another example of a method of correcting the first boundary of the first inner region of the skull according to an embodiment of the present application.

Referring to FIG. 53 , the image analysis apparatus 2000 according to an embodiment of the present application uses a region corresponding to a cervical vertebrae as a reference region RR to be the first of the first internal region IR1 of the skull. It may be implemented to correct a part of the boundary BD1.

As an example, the image analysis apparatus 2000 may acquire the region C1 related to the cervical vertebra. In this case, the image analysis apparatus 2000 may correct a part of the first boundary BD1 of the first internal region IR1 of the skull based on the region C1 related to the cervical vertebrae.

The image analysis apparatus 2000 may obtain a first feature region and a second feature region from the target image. For example, the first characteristic region may be a region corresponding to the anterior commissure (AC), and the second characteristic region may be a region corresponding to the posterior commissure (PC).

The image analysis apparatus 2000 may be implemented to calculate a first feature point from the first feature region and calculate a second feature point from the second feature region.

For example, the image analysis apparatus 2000 may include a center point of a region corresponding to the anterior commissure AC, which may be the first feature region, or a superior edge included in a boundary defining a region corresponding to the anterior commissure AC. edge), the first feature point may be calculated based on a region corresponding to an inferior edge. In addition, the image analysis apparatus 2000 includes a center point of a region corresponding to the posterior commissure (PC), which may be a second feature region, or an upper edge (Superior edge) included in a boundary defining a region corresponding to the posterior commissure (PC), The second feature point may be calculated based on an area corresponding to a lower edge or the like.

Also, the image analysis apparatus 2000 may acquire a reference plane based on the calculated first and second feature points.

For example, the image analysis apparatus 2000 may obtain a reference direction based on the center point of the anterior commissure AC and the central point of the posterior commissure PC on the sagittal plane. For example, the image analysis apparatus 2000 may obtain a reference direction connecting the central point of the anterior commissure AC and the central point of the posterior commissure PC. In this case, the image analysis apparatus 2000 may acquire the reference plane P2 adjacent to the reference region RR and parallel to the reference direction.

As another example, the image analysis apparatus 2000 may obtain the first plane P1 including the center point of the anterior commissure AC and the central point of the posterior commissure PC, and perpendicular to the sagittal plane P1. . In this case, the image analysis apparatus 2000 may acquire the reference plane P2 adjacent to the reference region RR and parallel to the first plane P1 .

The image analysis apparatus 2000 may obtain the second boundary BD2 by correcting a part of the first boundary B1 of the first inner region based on the reference plane P2 . For example, the image analysis apparatus 2000 corrects a portion of the first boundary BD1 positioned on the lower side of the reference plane P2 and the first positioned on the upper side of the reference plane P2 . The second boundary BD2 may be acquired based on the boundary BD1 .

More specifically, the image analysis apparatus 2000 replaces a portion of the first boundary BD1 located below the reference plane P2 with a reference line or a reference plane included in the reference plane P2, so that the first boundary ( BD1) can be implemented to modify

The image analysis apparatus 2000 may acquire the second boundary BD2 based on the modified first boundary, and may acquire the second internal region IR2 of the skull having the second boundary BD2.

53 shows an example of correcting a part of the first boundary BD1 of the first internal region IR of the skull only for the brain image related to the sagittal plane among the brain images, but the image analysis apparatus 2000 ) may perform a corresponding boundary correction operation on a brain image related to a coronal plane and a transverse plane among the brain images.

In addition, although it has been described in FIG. 53 that a part of the first boundary BD1 of the first internal region IR1 of the skull is corrected based on the region C1 of the cervical vertebrae, the C1 region of the cervical vertebrae must be corrected. It is not limited to modifying the first boundary BD1 based on . For example, the first boundary BD1 may be modified based on other regions of the cervical vertebrae, and the image analysis apparatus may modify the first boundary BD1 based on a region having prominent features in the brain image. (2000) can be implemented.

In addition, in FIG. 53 , the reference plane P2 is illustrated as a plane parallel to the first plane P1 obtained based on the anterior commissure (AC) and the posterior commissure (PC), but this is only an example, and image analysis The device 2000 may obtain a plane parallel to the transverse plane of the target image and including the region C1 as the reference plane P2, and based on the reference plane P2, the first boundary BD1 It can be implemented to modify.

In addition, according to FIGS. 51 to 53, although it is shown that the boundary is corrected for the two-dimensional image, this is only for convenience of explanation, and should be interpreted as obtaining the inner region by correcting the boundary with respect to the three-dimensional image. .

Referring again to FIG. 50 , the image analysis method according to an embodiment of the present application may include acquiring a morphological volume value of the second inner region ( S4430 ).

Specifically, in step S4430 of acquiring the morphological volume value of the second internal region, the image analysis apparatus 2000 may calculate the morphological volume value based on the acquired second internal region. In more detail, the image analysis apparatus 2000 may acquire a morphological volume value corresponding to the second internal region based on voxel data corresponding to the second internal region.

Also, the image analysis method according to an embodiment of the present application may include obtaining a morphological volume value of the target element ( S4440 ).

Specifically, in the step S4440 of obtaining the morphological volume value of the target element, the image analysis apparatus 2000 relates to the target element based on the region corresponding to the target element obtained in the image segmentation step S4300. Morphological volume values can be calculated. In more detail, the image analysis apparatus 2000 may be provided to obtain a morphological volume value corresponding to the target element based on voxel data corresponding to the target element.

In addition, the image analysis method according to an embodiment of the present application may include calculating a morphological index ( S4450 ). Specifically, in the step of calculating the morphological index (S4450), the image analysis apparatus 2000 corresponds to the morphological volume value corresponding to the second inner region obtained in the step S4430 and the target element obtained in the step S4450. A morphological index can be calculated based on the morphological volume value.

For example, the morphological index may be an index defined as a ratio of a morphological volume value corresponding to a target element to a morphological volume value corresponding to the second inner region.

일 수 있다. For example, morphological indicators =

can be

In addition, the calculated morphological index may provide meaningful information as a diagnostic auxiliary index related to brain diseases related to dementia, depression, stroke, and the like.

Although not shown in FIGS. 48 and 50 , each of the morphological volume value corresponding to the second inner region obtained in step S4430 and the morphological volume value corresponding to the target element obtained in step S4440 is related to the scan condition. The calibration may be based on a calibration parameter or a calibration parameter related to the position of the target element. In this case, the image analysis apparatus 2000 may be implemented to finally calculate the morphological index based on the morphologically corrected volume value corresponding to the second inner region and the morphologically corrected volume value corresponding to the target element. In this regard, it will be described later in detail with reference to FIGS. 58 to 67 .

Hereinafter, examples of information related to a morphological indicator output according to an embodiment of the present application and an operation of the image acquisition apparatus 2000 for outputting the information will be described with reference to FIGS. 54 to 57 . 54 to 57 are exemplary views illustrating information related to a morphological index output according to an embodiment of the present application.

Information related to output morphological indicators to be described later may be output through the output module 2050 or the output device 2600 of the image analysis apparatus 2000 according to an embodiment of the present application.

Referring to FIG. 54 , information related to a morphological index output from the image analysis apparatus 2000 according to an embodiment of the present application may include the total intracranial volume (ICV, Info. 1).

For example, the total intracranial volume (ICV, Info. 1) is a volume value obtained by the image analysis apparatus 2000 according to an embodiment of the present application based on voxel data corresponding to the internal region of the skull of the target image. can

At this time, as described above, the image analysis apparatus 2000 corrects a part of the first boundary of the first internal region of the skull obtained based on the segmentation result of the target image in order to define the internal region of the skull to define the second boundary. It is possible to obtain a second inner region having In this case, according to a preferred embodiment, the total intracranial volume ICV, Info. 1 may be a value obtained based on voxel data of the target image corresponding to the second inner region.

Referring back to FIG. 54 , the information related to the morphological index output from the image analysis apparatus 2000 according to an embodiment of the present application includes the volume of target elements (Info. 2), the morphological index (Info. 3), and the Percentile information (Info. 4) of morphological indicators may be included. In other words, the image analysis apparatus 2000 according to an embodiment of the present application includes a volume (Info. 2) of target elements, a morphological index (Info. 3), and a percentile (Info. 4) of the morphological index. Information related to academic indicators can be output.

The information output from the image analysis apparatus 2000 according to an embodiment of the present application may be information related to various regions located in the brain as the target element. In this case, the target element may be a region significantly related to a brain disease. For example, the target element may be a region related to the hippocampus, lateral ventricle, frontal lobe, temporal lobe, parietal lobe, occipital lobe, and amygdala associated with a brain disease such as dementia.

The image analysis apparatus 2000 according to an embodiment of the present application segments the target image into a plurality of brain regions including the above-described regions, and then, based on voxel data corresponding to each region, the volume of each of the target elements. It may be provided to obtain and output a value (volume value).

In addition, the image analysis apparatus 2000 according to an embodiment of the present application obtains the obtained volume values of the target elements, and calculates the volume values of the target elements in consideration of the scan condition in which the target image is captured or the location of the target element. can be corrected.

In this case, the volume Info. 2 of the target elements of FIG. 54 may be a corrected volume value.

The morphological index Info. 3 output from the image analysis apparatus 2000 according to an embodiment of the present application may be a ratio of a volume corresponding to the internal region of the skull to a volume corresponding to a target element. For example, referring to FIG. 54 , the morphological index (Info. 3) output from the image analysis apparatus 2000 is the volume value (or volume value) of target elements with respect to the total intracranial volume (ICV, Info. 1). , Info. 2).

However, this is only an example, and the image analysis apparatus 2000 may be provided to obtain and output morphological indicators related to shape, length, thickness, etc. related to any suitable brain disease without being limited to the volume.

The image analysis apparatus 2000 according to an embodiment of the present application may obtain a morphological index database, and may obtain a percentile (Info. 4) of the morphological index of the object based thereon. The percentile (Info. 4) of the morphological index output from the image analysis apparatus 2000 may be information related to the percentile of the morphological index (Info. 3) of the object located in the morphological indices of the comparison target group. In this case, the comparison target group may be a target group obtained in consideration of the gender or age of the target object of the target image.

Specifically, the image analysis apparatus 2000 may obtain percentile information of the morphological index of the object by using a morphological index database including one or more morphological indices obtained based on the comparison target brain image.

For example, the image analysis apparatus 2000 may obtain ICV data including one or more morphological values or morphological indices related to ICV obtained from the morphological index database, and obtain ICV data related to the obtained ICV based on the target image of the object. Percentiles on ICV data of morphological values or morphological indicators can be obtained.

For example, the image analysis apparatus 2000 may obtain information on the object related to the gender or age of the object of the target image, and obtain a morphological index database based on the information of the object. The morphological indices database may include one or more morphological indices obtained based on a brain image of a comparison target group having characteristics (gender, age, etc.) similar to information of the subject.

The image analysis apparatus 2000 may obtain percentile information of the morphological index of the object by using the obtained morphological index database.

For example, the image analysis apparatus 2000 obtains ICV data including at least one ICV-related morphological value or morphological index obtained based on a brain image of a male in his 60s, and obtains a brain image of a male in his 60s. A morphological value related to an ICV value obtained based on , or a percentile on ICV data of a morphological index may be obtained.

As another example, the image analysis apparatus 2000 according to an embodiment of the present application obtains information (eg, age or gender) of an object related to a target image, and from among a plurality of brain images stored in a database or memory 2020 The brain image of the comparison target group may be selected in consideration of the information of the target.

The image analysis apparatus 2000 generates one morphological value or morphological index related to the ICV corresponding to the target element through an operation similar to the operation of acquiring the morphological index from the target image from the brain images of the selected comparison target group. It is possible to obtain ICV data including the above.

The image analysis apparatus 2000 compares the morphological value or morphological index related to the ICV value obtained from the target image with the ICV data obtained from the brain images of the comparison target group, and a morphological value related to the ICV value obtained from the target image However, it is possible to calculate the percentile on the ICV data of the morphological indices.

Information related to the volume (Info. 2), the morphological index (Info. 3), and the percentile (Info. 4) of the target elements output from the image analysis apparatus 2000 according to an embodiment of the present application is Each of the left and right hemispheres may be individually acquired and output.

However, the information related to the morphological index shown in FIG. 54 is only an example, and it goes without saying that any appropriate information may be additionally obtained and output from the image analysis apparatus 2000 .

Information related to the volume (Info. 2), the morphological index (Info. 3), or the percentile (Info. 4) of the target elements output from the image analysis apparatus 2000 according to an embodiment of the present application is Information related to the morphological index may be processed and output by a statistical technique using any appropriate graph.

Hereinafter, exemplary contents in which information related to morphological indicators are processed and output by the image analysis apparatus 2000 according to an embodiment of the present application will be described with reference to FIGS. 55 to 57 .

Referring to FIG. 55 , information related to a morphological index output from the image analysis apparatus 2000 according to an embodiment of the application includes a percentile of the morphological index (Info. 5) and a graph of the percentile of the morphological index (Info. 6, 7) and the risk associated with brain disease (Info. 8).

The percentile (Info. 5) of the morphological index output from the image analysis apparatus 2000 according to an embodiment of the present application may be a numerical value obtained with respect to the target element located in the left hemisphere and the target element located in the right hemisphere.

For example, referring to FIG. 55 , percentile values of morphological indices obtained for the frontal lobe, temporal lobe, parietal lobe, occipital lobe, cingulate gyrus, hippocampus, and thalamus may be visually output. In this case, the percentiles of the morphological indices for the frontal lobe located in the left hemisphere and the percentiles of the morphological indexes for the frontal lobe located in the right hemisphere may be visually output together.

The graphs Info. 6 and 7 for percentiles of the morphological indices output from the image analysis apparatus 2000 according to an embodiment of the present application are the morphological indices Info. 3) may be generated based on a percentile of the morphological index of the subject calculated by comparing the morphological index obtained from the brain images of the comparison target group.

In this case, the graph may be provided to visually display the percentile of the morphological index, and may include a ruler (Info. 7) related to the percentile indicating where the percentile of the morphological index of the object is distributed.

In addition, the graph may be provided by displaying the percentiles of the morphological indices for the left hemisphere and the percentiles of the morphological indices for the right hemisphere to be distinguished.

In the information related to the risk (Info. 8) related to brain disease output from the image analysis apparatus 2000 according to an embodiment of the present application, the risk may be calculated based on the percentile of the morphological index of the subject, and the risk may be may be provided to be visually differently output according to the .

Specifically, the image analysis apparatus 2000 may be provided such that a predetermined threshold value is set with respect to the percentile of the morphological index in relation to the risk (Info. 8) associated with the brain disease. In this case, the predetermined threshold value may be differently preset according to the degree of risk (eg, caution, danger, normal).

In this case, when the percentile of the morphological index of the object is smaller than a predetermined first threshold, the image analysis apparatus 2000 may be implemented to visually output information indicating that the object may have a brain disease and is at risk. For example, referring to FIG. 55 , the image analysis apparatus 2000 may determine that the percentile of the morphological indices of the hippocampus located in the left and right hemispheres of the object is smaller than a predetermined first threshold value, so it may be determined to be dangerous. It may be implemented to output the percentile of the morphological index related to the hippocampus as a first color (eg, red) based on the determination result.

Also, when the percentile of the morphological index of the object is greater than the first predetermined threshold and less than the second predetermined threshold, the image analysis apparatus 2000 indicates that there is a certain possibility that the object has a brain disease and requires attention. It may be implemented to visually output information. For example, when the percentile of the morphological index of a specific target element located in the left and right hemispheres of the object is greater than a first predetermined threshold and smaller than a second predetermined threshold, the image analysis apparatus 2000 may may be implemented to determine that attention is necessary, and may be implemented to output the percentile of the morphological index of a specific target element as a second color (eg, orange) based on the determination that attention is required.

However, outputting in different colors according to the risk level described above is only an example, and the image analysis apparatus 2000 may be implemented to provide information related to the risk level related to a brain disease using any suitable method without being limited thereto. of course there is

The image analysis apparatus 2000 according to an embodiment of the present application may be implemented to provide information related to the morphological indices of the object obtained from the target image and the morphological indices obtained from the brain images of the comparison target group. .

Specifically, as described above with reference to FIG. 55 , the image analysis apparatus 2000 determines the morphological index of the comparison target group from the morphological index database based on the information of the object including the object's age or gender of the target image. can be obtained

In this case, the image analysis apparatus 2000 according to an embodiment of the present application may be implemented to output the morphological index of the object and the morphological index of the comparison target group in the form of a visual graph according to information on the object. In this case, the morphological index of the comparison target group according to the information of the object reflects the morphological index of the comparison target group corresponding to the average value of the morphological indices of the comparison target group and predetermined threshold values related to the risk for brain disease can be implemented to be output.

See FIG. 56 . 56 is a graph illustrating an example of information related to a morphological index output by the image analysis apparatus 2000 according to an embodiment of the present application.

For example, the morphological index output by the image analysis apparatus 2000 according to an embodiment of the present application is a morphological index (Info. 9) of the object according to information (eg, age or gender) of the object related to the target image. ) can be output as

In this case, the image analysis apparatus 2000 according to an embodiment of the present application provides the morphological indices (Info. 9) of the subject and the morphological indices (Info. 9) of the target group in order to provide information related to the subject's brain disease. .10, Info.11) may be additionally output.

For example, as described above, the image analysis apparatus 2000 includes morphological index information of a comparison target group in their 70s and 80s from a morphological index database in consideration of the subject's age (eg, 79 years of age in FIG. 56 ) as described above. can be obtained In this case, the image analysis apparatus 2000 may be implemented to obtain and output the average value (Info. 10) of the morphological index information of the comparison target group obtained in consideration of the age.

As another example, as described above, the image analysis apparatus 2000 may set threshold values related to morphological indices of comparison target groups to provide information on the risk associated with a brain disease. For example, when the percentile of the morphological index of the object is less than 5%, the image analysis apparatus 2000 may be implemented such that a threshold is set in advance to determine that it is dangerous. Alternatively, when the percentile of the morphological index of the object is 5% or more and less than 10%, the image analysis apparatus 2000 may be implemented to set a threshold in advance to determine that attention is required.

At this time, the image analysis apparatus 2000 according to an embodiment of the present application may be implemented to output morphological indicators (Info. 11) of comparison target groups corresponding to a preset threshold in relation to the degree of risk for brain disease. can In this case, the morphological indices Info. 11 of the comparison target groups may be output according to information on the object, such as an age of the object.

The image analysis apparatus 2000 according to an embodiment of the present application provides information (Info. 12) may be provided.

For example, when the percentile of the morphological index of the object is less than 5%, the image analysis apparatus 2000 may be implemented such that a threshold is set in advance to determine that it is dangerous. In this case, when the percentile of the morphological index of the object is less than 5%, the image analysis apparatus 2000 may be implemented to visually output information indicating that the object may be in a dangerous state for a specific brain disease. For example, referring to FIG. 56 , the image analysis apparatus 2000 may be implemented so that the morphological index (Info. 9) of the object is displayed and output in a first color (eg, red).

Alternatively, when the percentile of the morphological index of the object is 5% or more and less than 10%, the image analysis apparatus 2000 may be implemented to set a threshold in advance to determine that attention is required. In this case, when the percentile of the morphological index of the object is 5% or more and less than 10%, the image analysis apparatus 2000 is implemented to visually output information indicating that the object may be in a state requiring attention for a specific brain disease. can be For example, the image analysis apparatus 2000 may be implemented so that the morphological indicator Info. 9 of the object is displayed and output in the second color (eg, orange).

However, the above-described content is merely an example and is not limited thereto. That is, the image analysis apparatus 2000 may be implemented to output information related to the subject's risk for brain disease using any suitable method.

The image analysis apparatus 2000 according to an embodiment of the present application visually outputs information related to three-dimensional brain information in order to provide information on the three-dimensional structure of the target element or where the target element is located in the brain. may be provided. Hereinafter, an example of information related to a morphological index output according to an embodiment of the present application and an operation of the image acquisition apparatus 2000 for outputting the information will be described with reference to FIG. 57 .

For example, the image analysis apparatus 2000 may additionally acquire data related to the 3D brain template, based on the data related to the 3D brain template and the region corresponding to the target element obtained by segmentation of the target image. Accordingly, information on the three-dimensional structure of the target element or the location of the target element inside the brain can be output in a visual graphic form (Info. 13).

In addition, the image analysis apparatus 2000 according to an embodiment of the present application provides a morphology related to a target element based on the percentile of the data related to the three-dimensional brain template and the morphological index related to the target element calculated as described above. It can be implemented to overlay (Info. 14) information on the percentile of the index on a three-dimensional brain template and output it. For example, the image analysis apparatus 2000 according to an embodiment of the present application overlays different colors on a three-dimensional brain template according to percentiles of morphological indices related to the target element, thereby providing information related to the degree of volume atrophy of the target element. It can be implemented to provide

In addition, the image analysis apparatus 2000 according to an embodiment of the present application receives an input for outputting information related to a detailed morphological index from a user through an input module, and outputs an output corresponding to the user's input to the output module It can be implemented to output to the user through For example, when a user's input for a specific brain region of the 3D brain template shown in FIG. 57 is obtained, the image analysis apparatus 2000 provides information related to a morphological index for a specific brain region corresponding to the obtained input. data (eg, information shown in FIGS. 54 to 56 ) may be output.

However, the information related to the morphological indices shown in FIG. 57 is merely an example, and the image analysis apparatus 2000 according to an embodiment of the present application may use any suitable method for percentiles of the morphological indices related to the target element. It may be implemented to provide information in three dimensions.

Also, the image analysis apparatus 2000 according to an embodiment of the present application may be implemented to output information related to a morphological index related to an object according to a temporal variable. For example, the image analysis apparatus 2000 obtains information related to morphological indices obtained from the object at the first time point (eg, information shown in FIGS. 54 to 57 ) and the information obtained from the object at the second time point. It may be implemented to output information (eg, information shown in FIGS. 54 to 57 ) related to the morphological indices. In particular, although it is important to compare the morphological indices with the comparison group in relation to brain diseases, it may also be important to analyze the morphological indices of the subject according to temporal variables.

The image analysis apparatus 2000 according to an embodiment of the present application may acquire information related to the morphological index at the first time point through the image analysis operations described above from the brain image obtained from the object at the first time point, , information related to the morphological index at the second time point may be acquired from the brain image obtained from the object at the second time point different from the first time point through the image analysis operations described above. In this case, the image analysis apparatus 2000 may be implemented to output the information related to the morphological index at the first time point and the information related to the morphological index at the second time point to the user through various statistical techniques.

However, the information related to the morphological indices described above with reference to FIGS. 54 to 57 is only an example, and any appropriate type of brain disease-related information may be output to the user by any suitable method.

An index that quantifies the degree of brain contraction in relation to brain-related diseases, particularly dementia, is being used as an auxiliary index in the diagnosis of dementia. In this case, the numerical value related to the degree of contraction of the brain may be measured differently depending on the scan condition in which the brain image is taken or the location of the target element. Here, the position of the target element may mean a position where the target element is distributed inside the skull.

For example, the morphological index obtained for the same object, that is, the numerical value related to the degree of contraction of the brain, includes the strength of the magnetic field of the image acquisition device from which the brain image was taken, the manufacturer of the image acquisition device, or setting parameters of the image acquisition device may depend on the scanning conditions. As a specific example, the morphological indices obtained for the same object may be measured differently according to a variable called the strength of the magnetic field.

For another example, a morphological index obtained for the same object, that is, a numerical value related to the degree of contraction of the target element, may be measured differently depending on the location of the target element in the skull.

Therefore, it is required to correct the morphological index according to the scan condition in which the brain image was taken or the location of the target element.

The image analysis apparatus 2000 according to an embodiment of the present application is implemented to perform an operation of correcting a morphological value or a morphological index of a target element in consideration of a scan condition in which a brain image is captured or a location of a target element can be

Accordingly, the image analysis apparatus 2000 according to an embodiment of the present application may more accurately acquire a morphological index related to a degree of brain contraction related to a brain disease, and may provide a more objective morphological indicator related to a brain disease diagnosis. It is possible to provide an advantageous effect that the auxiliary information can be provided to the user.

Hereinafter, a method of correcting a morphological value implemented by the image analysis apparatus 2000 according to an embodiment of the present application will be described in detail with reference to FIGS. 58 to 67 . Hereinafter, the morphological values related to the volume will be mainly described, but the present invention is not limited thereto, and any appropriate morphological characteristics such as the shape, thickness, and length of the target site that can be indicators of brain diseases other than the volume are equally corrected. method can be applied.

See Figure 58. 58 is a flowchart illustrating an operation of a morphological numerical correction method implemented by the image analysis apparatus 2000 according to an embodiment of the present application. Specifically, FIG. 58 is a flowchart illustrating an operation of correcting a brain-related morphological value or a brain-related morphological index implemented by the image analysis apparatus 2000 according to an embodiment of the present application.

A method of correcting a morphological value according to an embodiment of the present application includes: obtaining a target image and a scan condition of the target image (S5100); It may include the step of segmenting the target image (S5200) and the step of correcting the brain morphological index (S5300).

In operation S5100 of acquiring the target image and the scan condition of the target image, the image analysis apparatus 2000 may acquire the target image from the image acquisition apparatus 1000 . The target image may mean encompassing the brain image to be analyzed by the image analysis apparatus 2000 . In other words, the target image may include the target brain image from which the morphological value or morphological index of the target element is to be calculated from the image analysis apparatus 2000 .

In addition, in step S5100 of acquiring the target image and the scan condition of the target image, the image analysis apparatus 2000 may obtain information related to the scan condition in which the target image is captured from the image acquisition apparatus 1000 . As described above, the information related to the scan condition in which the target image is captured may be structured as metadata for the target image, and may be acquired together by acquiring the target image from the image acquiring apparatus 1000 . Alternatively, information related to a scan condition in which the target image is captured may be acquired from an arbitrary external device separately from the target image.

In the segmentation of the target image ( S5200 ), the image analysis apparatus 2000 may perform segmentation of the target image according to the image segmentation operation described above with reference to FIGS. 7 to 17 .

According to an example, at least one target element included in the target image may be acquired through the segmentation operation of the image analysis apparatus 2000 .

Also, an inner region of the skull, which will be described later, may be acquired by the segmentation operation of the image analysis apparatus 2000 . In order to obtain the internal region of the skull, the image analysis apparatus 2000 modifies a part of the first boundary of the first internal region of the skull, which was described above with reference to FIGS. 48 and 57 , so that the second boundary of the skull has a second boundary. It may be implemented to perform an operation for acquiring the inner region.

Also, the neural network model provided to the image analysis apparatus 2000 may be trained to segment a brain image according to a scan condition. For example, a first neural network model learned from training data including a brain image obtained under the first scan condition may be provided to segment the target image obtained under the first scan condition. As another example, a second neural network model learned from training data including a brain image obtained under the second scan condition may be provided to segment the target image obtained under the second scan condition.

Referring to FIG. 58, although the pre-processing of the target image and the alignment of the target image are not shown, the operations of the image analysis apparatus 2000 in the pre-processing of the image and the alignment of the image (S4200) described in relation to FIG. 48 are The same can be applied.

Hereinafter, an embodiment of the step ( S5300 ) of correcting the brain morphological index will be described in more detail with reference to FIG. 59 . 59 is a view showing a flowchart of an image analysis method according to an embodiment of the present application, and in detail, the step S52600 of FIG. 58 is subdivided.

Referring to FIG. 59 , the step of correcting the brain morphological index according to an embodiment of the present application (S5300), obtaining the morphological value of the target element based on the segmentation result (S5310), obtaining the correction parameter Further comprising the step of (S5320), obtaining a morphological correction value based on the morphological value and the correction parameter (S5330), and obtaining and outputting a brain morphological index based on the morphological correction value (S5340) can do.

In operation S5310 of acquiring the morphological value of the target element based on the segmentation result, the image analysis apparatus 2000 may perform the segmentation operation on the basis of voxel data of the target image related to the region corresponding to the target element. A morphological value related to the target element may be obtained. Alternatively, the image analysis apparatus 2000 may acquire a morphological value related to the internal region of the skull based on voxel data of the target image related to the internal region of the skull obtained by the segmentation operation.

In addition, in the step of obtaining the morphological value of the target element based on the segmentation result ( S5310 ), the image analysis apparatus 2000 , based on the morphological value related to the target element and the morphological value related to the inner region of the skull, A morphological index related to the target element may be obtained.

In this case, the morphological value or brain morphological index of the target element may be a numerical value related to a morphological character including the volume, thickness, length, and shape of the target element.

In the step of acquiring the correction parameter ( S5320 ), the image analysis apparatus 2000 obtains the correction parameter from the parameter acquisition apparatus 2400 and corrects the morphological value related to the target element or the morphological value related to the inner region of the skull. It can be implemented to perform an operation.

Alternatively, in the step of acquiring the correction parameter ( S5320 ), the image analysis apparatus 2000 may be implemented to acquire the correction parameter from the parameter acquisition apparatus 2400 and correct the morphological index related to the target element. have. That is, according to FIG. 59, although it is illustrated that the correction parameter is applied to the morphological value of the target element and corrected, this is only an example and is not limited thereto. The image analysis apparatus 2000 may be implemented to apply and output a correction parameter to the brain morphological index calculated based on the value.

In this case, the parameter obtaining apparatus 2400 may obtain the correction parameters through correlation analysis related to a morphological value according to a scan condition or a location of a target element. The operation of the correction parameter obtaining apparatus 2400 in relation to the step of obtaining the correction parameter ( S5320 ) will be described in detail later with reference to FIGS. 60 to 65 .

In step S5330 of obtaining a morphological correction value based on the morphological value and the correction parameter, the image analysis apparatus 2000 includes the morphological value corresponding to the target element and the correction obtained from the correction parameter acquiring apparatus 2400 . It may be implemented to perform an operation of obtaining a morphological correction value related to the target element based on the parameter.

The correction parameter may be a parameter for approximating or estimating the morphological value related to the target element obtained under the first scan condition to the morphological value obtained under the second scan condition.

For example, the image analysis apparatus 2000 obtains a morphological correction value related to the target element based on a morphological value corresponding to the target element and a correction parameter included in a transformation function obtained from the correction parameter acquiring apparatus 2400 . can do.

In this case, the transform function may be an arbitrary nth-order function, and the correction parameter may mean a coefficient included in the transform function.

For example, the correction parameter is obtained from a transformation function obtained through correlation analysis between a first morphological value related to the target element obtained under the first scan condition and a second morphological value related to the target element obtained under the second scan condition. can be For example, if the first-order function obtained through the correlation analysis between the first morphological value and the second morphological value is y=a*x+b, a and b are the values of the target element obtained under the first scan condition. It may be a correction parameter for approximating or estimating the morphological value to the morphological value obtained under the second scan condition. The image analysis apparatus 2000 multiplies a morphological value related to a target element obtained from a target image obtained under the first scan condition by a correction parameter (a) and adds the correction parameter (b) to obtain under the second scan condition It can be approximated or estimated as a morphological value.

Meanwhile, the difference between the morphological measurement value of the target element and the morphological true value (or morphological reference value) may be different according to the location of the target element obtained from the target image. The image analysis apparatus 2000 according to an embodiment of the present application may correct the morphological measurement value of the target element in consideration of the location of the target element.

For example, the image analysis apparatus 2000 performs the first correction from the correction parameter acquisition apparatus 2400 in order to approximate the morphological measurement value of the first brain element to the morphological true value (or morphological reference value) of the first brain element. parameters can be obtained.

On the other hand, the image analysis apparatus 2000 obtains the second correction parameter from the correction parameter acquisition apparatus 2400 in order to approximate the morphological measurement value of the second brain element to the morphological true value (or morphological reference value) of the second brain element. can be obtained

Also, the image analysis apparatus 2000 according to an embodiment of the present application may obtain a correction parameter in consideration of both a scan condition and a variable related to a location of a target element.

For example, the image analysis apparatus 2000 may approximate a morphological value related to the first brain element of the target image obtained under the first scan condition to the morphological value obtained under the second scan condition using the first correction parameter. have. On the other hand, the image analysis apparatus 2000 may approximate the morphological value related to the second brain element of the target image obtained under the first scan condition to the morphological value obtained under the second scan condition using the second correction parameter. have.

For example, if the first-order function obtained through correlation analysis between the first morphological value and the second morphological value in relation to the first brain element is y=a1*x+b1, then a1 and b1 are the first scans It may be a correction parameter for approximating the morphological value of the first brain element obtained under the condition to the morphological value obtained under the second scan condition. On the other hand, if the first-order function obtained through the correlation analysis between the first morphological value and the second morphological value in relation to the second brain element is y=a2*x+b2, a2 and b2 under the first scan condition It may be a correction parameter for approximating the obtained morphological value of the second brain element to the obtained morphological value under the second scan condition.

The image analysis apparatus 2000 multiplies the morphological value related to the first brain element obtained from the target image obtained under the first scan condition by the correction parameter (a1) and adds the correction parameter (b1) to the second scan condition It can be approximated by the morphological values obtained under On the other hand, with respect to the second brain element related to the target image obtained under the first scan condition, the image analysis apparatus 2000 multiplies the morphological value related to the second brain element by the correction parameter (a2), and sets the correction parameter (b2) By adding, it can be approximated to the morphological value obtained under the second scan condition.

However, the above-described conversion function is merely an example, and by obtaining any suitable function, the correction of the morphological value in consideration of the scan condition may be performed.

The image analysis apparatus 2000 according to an embodiment of the present application may approximate a morphological value to a value obtained under a consistent scan condition based on values obtained under various scan conditions, and through this, a consistent brain disease-related It can provide diagnostic auxiliary indicators.

In addition, since the image analysis apparatus 2000 according to an embodiment of the present application may perform correction by differently acquiring correction parameters in consideration of the location of the target element, it is possible to provide a more accurate diagnostic auxiliary indicator related to brain disease. can

In the step of acquiring and outputting the brain morphological index based on the morphological correction value ( S5340 ), the image analysis apparatus 2000 includes a first morphological correction value related to the target element and a second shape related to the internal region of the skull It may be implemented to calculate and output a brain morphology index based on a phylogenetic correction value.

As an example, the brain morphological index may be defined as a first morphological correction value with respect to the second morphological correction value. However, this is only an example, and a brain morphological index may be defined as any suitable diagnostic auxiliary index related to a brain disease.

Meanwhile, there may be a plurality of target elements. In other words, the target element may include a first brain element and a second brain element.

In this case, the image analysis apparatus 2000 may segment the target image to obtain a first region corresponding to the first brain element, a second region corresponding to the second brain element, and a skull region. Also, the image analysis apparatus 2000 may acquire the inner region of the skull based on the skull region.

In this case, as described above, the image analysis apparatus 2000 may calculate the first brain element and the first corrected morphological value based on the correction parameter and the voxel data of the first region, and the correction parameter and the voxel of the second region A second brain element and a second corrected morphological value may be calculated based on the data. Also, the image analysis apparatus 2000 may acquire a reference morphological value related to the internal region of the skull based on voxel data corresponding to the internal region of the skull, and this value is used as a reference for calculating the brain morphological index. can be In this case, the correction parameter related to the first brain element and the correction parameter related to the second brain element may be different. This will be described later with reference to FIG. 64 .

The image analysis apparatus 2000 may calculate a first brain morphological index related to the first brain element based on the first corrected morphological value and the reference morphological value. Also, the image analysis apparatus 2000 may calculate a second brain morphological index related to the second brain element based on the second corrected morphological value and the reference morphological value.

The image analysis apparatus 2000 according to an embodiment of the present application may determine a correction parameter for correcting a morphological value obtained from the target image in consideration of characteristics including the age and gender of the object of the target image.

For example, when the target image is obtained from a first object having a first characteristic related to gender and age, the image analysis apparatus 2000 may set a correction parameter for correcting a morphological value obtained from the target image, the first It may be determined by the correction parameters obtained from the first brain image and the second brain image obtained from the second object having the characteristics. For example, the image analysis apparatus 2000 may use a correction parameter obtained from a second object having a similar age and the same sex as that of the first object of the target image to correct a morphological value or morphological index obtained from the target image. can be decided with

In the step of obtaining and outputting the brain morphological index based on the morphological correction value ( S5340 ), the image analysis apparatus 2000 outputs information related to the obtained brain morphological index to the output module 2050 or the output device 2600 . It can output through the output module 2650 included in the . Specifically, the image analysis apparatus 2000 may be implemented to similarly and visually output information related to FIGS. 54 to 57 .

On the other hand, according to FIG. 59 , although it is illustrated that the brain morphological index related to the target element is output after correcting the morphological value related to the target element, the brain morphological index is calculated based on the morphological value related to the target element. The image analysis apparatus 2000 and the parameter acquisition apparatus 2400 may also be implemented in a manner of correcting and outputting the brain morphological index in consideration of the correction parameters after calculation.

According to the present embodiment, the image analysis apparatus 2000 determines a target morphological value corresponding to a region related to a target element of a target image photographed under the first scan condition under the second scan condition or the third scan condition other than the first scan condition. It can be calculated as a morphological estimate of the target element under scan conditions.

Accordingly, the image analysis apparatus 2000 according to the present embodiment can freely correct the target morphological value into a morphological estimate value corresponding to a more accurate scan condition, and thus output morphological index information with improved accuracy. have.

Hereinafter, an operation in which the correction parameter obtaining apparatus 2400 according to an embodiment of the present application acquires the correction parameter will be described in detail with reference to FIGS. 60 to 65 .

60 is a flowchart illustrating an operation of a method of obtaining a correction parameter implemented by the apparatus 2400 for obtaining a correction parameter according to an embodiment of the present application. A correction parameter to be described later may be obtained by the correction parameter obtaining apparatus 2400 according to an embodiment of the present application. In other words, the operation of acquiring the correction parameter illustrated in FIG. 60 may be implemented by the correction parameter acquiring apparatus 2400 .

Referring to FIG. 60 , the method of obtaining a correction parameter according to an embodiment of the present application includes obtaining a plurality of brain image data sets ( S6100 ), and correlation of morphological values according to scan conditions of at least one data set. It may include a step of analyzing (S6200) and obtaining a correction parameter based on the correlation analysis result (S6300).

In the step of acquiring a plurality of brain image data sets ( S6100 ), the correction parameter acquiring apparatus 2400 according to an embodiment of the present application may acquire a plurality of brain image data sets. In this case, the correction parameter obtaining apparatus 2400 may obtain a plurality of brain image data sets from the image obtaining apparatus 1000 or any external device.

See FIG. 61 . 61 is an exemplary diagram for explaining information included in a plurality of brain image data sets according to an embodiment of the present application.

The plurality of brain image data sets may include information about a morphological value or a morphological index of a region corresponding to a target element related to a brain image.

In addition, information on a plurality of brain image data sets, information on the scan condition or the location of the target element under which the brain image was obtained, and information on morphological values or morphological indicators obtained in relation to the scan condition or the location of the target element may include

For example, the first data set includes information related to a first morphological value (or a first morphological index, V1) calculated from the first brain image IM1 acquired under the first scan condition and information related to the second scan condition under the second scan condition. Information related to the second morphological value (or the second morphological index, V2) calculated from the acquired second brain image IM2 may be included. Also, the first data set may further include information related to a third morphological value (or a third morphological index, V3) calculated from the third brain image IM3 acquired under the third scan condition.

In addition, the plurality of brain image data sets may further include various types of brain data similar in shape to the first data set. For example, the n-th data set includes information related to a first morphological value (or a first morphological index, V1n) calculated from a first brain image IM1n acquired under the first scan condition and the second scan condition. information related to the second morphological value (or the second morphological index, V2n) calculated from the second brain image IM2n obtained under Also, the n-th data set may further include information related to a third morphological value (or a third morphological index, V3n) calculated from the third brain image IM3n acquired under the third scan condition.

Preferably, the plurality of brain image data sets may include morphological values (or morphological indicators) according to scan conditions of brain images acquired for the same object. For example, referring to FIG. 61 , the first brain image IM1 , the second brain image IM2 , and the third brain image IM3 included in the first data set are brain images obtained from the first object. can Also, the first brain image IM1n, the second brain image IM2n, and the third brain image IM3n included in the n-th data set may be brain images obtained from the n-th object.

In this case, the first to third scan conditions are the strength of the magnetic field (or the strength of the magnetic field) related to the resolution of the image of the image acquisition device, and setting parameters related to the shape of the generated magnetic field that can be set in the image acquisition device , may be conditions related to the type of the manufacturer of the image acquisition device or a combination thereof.

Meanwhile, the correction parameter acquisition apparatus 2400 according to an embodiment of the present application may acquire the correction parameter by using a data set in consideration of the characteristics of the object. The correction parameter obtaining apparatus 2400 may obtain the correction parameter by using a data set obtained from objects having one or more identical characteristics (eg, age or gender). The correction parameter obtaining apparatus 2400 is configured to apply one or more correction parameters to the image analysis apparatus ( 2000) can be transferred.

Referring back to FIG. 60 , the method of obtaining a correction parameter according to an embodiment of the present application may include analyzing a correlation between a scan condition of at least one data set and a morphological value ( S6200 ).

In the step of analyzing the correlation between the scan condition of at least one data set and the morphological value ( S6200 ), the correction parameter obtaining apparatus 2400 according to an embodiment of the present application provides at least one of a plurality of brain image data sets. It may be implemented to analyze the correlation between morphological values (or morphological indicators) according to scan conditions included in the above data sets. In this case, the correlation analysis implemented by the correction parameter obtaining apparatus 2400 according to an embodiment of the present application may be implemented using various statistical techniques such as linear analysis, nonlinear analysis, and regression analysis.

The method of obtaining a correction parameter according to an embodiment of the present application may include obtaining a correction parameter based on a correlation analysis result ( S6300 ).

Specifically, the correction parameter obtaining apparatus 2400 may include a first morphological value (V1i) obtained under a first scan condition from an i-th data set included in at least one data set among a plurality of brain image data sets. , a second morphological value V2i obtained under the second scan condition and a third morphological value V3i obtained under the third scan condition may be obtained.

In addition, the correction parameter obtaining apparatus 2400 may include, from the i-th data set as well as the arbitrary j-th data set, the first morphological value V1j obtained under the first scan condition, and the second form obtained under the second scan condition. The morphological value V2j and the third morphological value V3j obtained under the third scan condition may be obtained.

In this case, the correction parameter obtaining apparatus 2400 according to an embodiment of the present application may analyze the correlation between the first morphological value and the second morphological value obtained from the i and j data sets, and the correlation analysis Based on the result, a correction parameter between the morphological value obtained under the first scan condition and the morphological value obtained under the second scan condition may be obtained.

Also, the correction parameter obtaining apparatus 2400 may analyze a correlation between the second morphological value and the third morphological value obtained from the i and j data sets, and based on the correlation analysis result, the second scan condition A correction parameter between the morphological value obtained under the condition and the morphological value obtained under the third scan condition may be obtained.

Also, the correction parameter obtaining apparatus 2400 may analyze a correlation between the first morphological value and the third morphological value obtained from the i and j data sets, and based on the correlation analysis result, the first scan condition A correction parameter between the morphological value obtained under the condition and the morphological value obtained under the third scan condition may be obtained.

For example, the correction parameter obtaining apparatus 2400 may perform a correlation analysis based on the first morphological value obtained under the first scan condition and the second morphological value obtained under the second scan condition, and the shape obtained under the first scan condition A correction parameter for converting a morphological value into a morphological value measured under the second scan condition may be obtained. Alternatively, a correction parameter for converting a morphological value obtained under the second scan condition into a morphological value measured under the first scan condition may be acquired.

For example, the correction parameter is obtained from a transformation function obtained through correlation analysis between a first morphological value related to the target element obtained under the first scan condition and a second morphological value related to the target element obtained under the second scan condition. can be For example, if the first-order function obtained through the correlation analysis between the first morphological value and the second morphological value is y=a*x+b, a and b are the values of the target element obtained under the first scan condition. It may be a correction parameter for approximating the morphological value to the morphological value obtained under the second scan condition.

Also, the correction parameter obtaining apparatus 2400 approximates the morphological value of the target element obtained under the second scan condition to the morphological value of the target element obtained under the third scan condition through correlation analysis similarly to the above-mentioned. a correction parameter for approximating the morphological value of the target element obtained under the first scan condition to the morphological value of the target element obtained under the third scan condition, and the shape of the target element obtained under the second scan condition A correction parameter for approximating the morphological value to the morphological value of the target element obtained under the first scan condition or the morphological value of the target element obtained under the third scan condition is the morphological value of the target element obtained under the first scan condition A correction parameter for approximation can be obtained.

Reference is made to FIGS. 62 and 63 .

62 is a graph illustrating an example of correlation analysis for acquiring a correction parameter of the correction parameter acquiring apparatus 2400 according to an embodiment of the present application.

63 is a graph illustrating another example of correlation analysis for obtaining a correction parameter of the apparatus 2400 for obtaining a correction parameter according to an embodiment of the present application.

Referring to FIG. 62 , the correction parameter obtaining apparatus 2400 includes a morphological value (eg, volume of a target element) obtained under a first magnetic field strength (eg, 1.5T) and a second magnetic field strength ( For example, it is possible to obtain a morphological value (eg, volume of a target element) obtained under 3T or 8T). At this time, the correction parameter acquisition device 2400 compares the morphological values obtained under the first magnetic field strength (eg, 1.5T) with the correlation analysis method using any suitable statistical technique such as the linear analysis shown in FIG. 62 . A correction parameter between the morphological values obtained under the second magnetic field strength (eg, 3T or 8T) may be obtained. The correction parameter obtaining apparatus 2400 may obtain a correction parameter for correcting the morphological value obtained under the first magnetic field strength to the morphological value obtained under the second magnetic field strength.

Referring to FIG. 63 , the correction parameter obtaining device 2400 includes a morphological value (eg, volume of a target element) obtained from an image device of a first manufacturer and a morphological value obtained from an image device of a second manufacturer from the j-th data set. A morphological figure (eg, the volume of a target element) may be obtained. At this time, the correction parameter acquisition device 2400 compares the morphological values obtained from the image device of the first manufacturer and the A correction parameter between the morphological values obtained from the imaging device may be obtained. The correction parameter obtaining device 2400 is a correction parameter for correcting the morphological value obtained from the brain image obtained from the image device of the first manufacturer to the morphological value obtained from the brain image obtained from the image device of the second manufacturer. can be obtained

However, the correlation analysis described in FIGS. 63 and 64 is only an example, and any appropriate statistical technique such as nonlinear analysis or regression analysis may be used to obtain correction parameters between morphological values according to scan conditions. .

In addition, FIGS. 63 and 64 exemplify a method of obtaining a correction parameter between a morphological value in consideration of a magnetic field strength and a manufacturer's variable as a scan condition, respectively, but in a similar manner, considering the scan condition related to the setting parameter of the imaging device It is obvious that calibration parameters can be obtained.

In addition, correction parameters between morphological values according to scan conditions may be obtained by considering at least two scan conditions among scan conditions related to magnetic field strength, scan conditions related to manufacturers, and scan conditions related to setting parameters of the imaging device. am.

Reference is again made to FIG. 60 . The method of obtaining a correction parameter according to an embodiment of the present application may include analyzing a correlation between morphological values according to scan conditions of at least one data set ( S6200 ). In this case, the correction parameter obtaining apparatus 2400 may be implemented to obtain the correction parameter by performing correlation analysis in consideration of information related to the location of the target element other than the scan condition.

In other words, even if the scan conditions are the same, parameters for correcting the morphological index may be different depending on the location of the target element. It may be implemented to obtain a correction parameter for correcting the morphological index in consideration.

Specifically, the difference between the morphological measurement value and the morphological true value (or morphological reference value) of the target element may be different according to the location of the target element obtained from the target image. The image analysis apparatus 2000 according to an embodiment of the present application may correct the morphological measurement value of the target element in consideration of the location of the target element.

For example, the image analysis apparatus 2000 performs the first correction from the correction parameter acquisition apparatus 2400 in order to approximate the morphological measurement value of the first brain element to the morphological true value (or morphological reference value) of the first brain element. parameters can be obtained. On the other hand, the image analysis apparatus 2000 acquires the second correction parameter from the correction parameter acquisition apparatus 2400 in order to approximate the morphological measurement value of the second brain element to the morphological true value (or morphological reference value) of the second brain element. can do.

For another example, the image analysis apparatus 2000 may obtain a first from the correction parameter obtaining apparatus 2400 in order to approximate the morphological measurement value of the first brain element to the morphological true value (or morphological reference value) of the first brain element. Calibration parameters can be obtained. On the other hand, the image analysis apparatus 2000 may determine that the morphological measurement value of the second brain element is substantially the same as the morphological true value (or morphological reference value) of the second brain element, and is obtained from the correction parameter obtaining apparatus 2400 . It may be implemented not to acquire the second correction parameter.

In other words, the image analysis apparatus 2000 may be implemented to differently correct whether the morphological measurement value is corrected according to the location of the target element.

In this case, an arbitrary threshold value for determining whether the morphological measurement value of the brain element is substantially the same as the true morphological value of the brain element may be set.

Hereinafter, as information related to the location of the target element other than the scan condition is further considered, the content that is changed or added with respect to the method of obtaining the correction parameter according to an embodiment of the present application will be mainly described.

In the step of acquiring the plurality of brain image data sets ( S6100 ), the correction parameter acquiring device 2400 according to an embodiment of the present application obtains the plurality of brain image data sets from the image acquiring device 1000 or any external device. can be obtained In this case, the correction parameter obtaining apparatus 2400 may obtain a plurality of brain image data sets from a database or any external device.

See FIG. 64 . 64 is an exemplary diagram for explaining information included in a plurality of brain image data sets according to an embodiment of the present application.

The plurality of brain image data sets may include information on a morphological value or a morphological index according to a scan condition in which a brain image is obtained and a location of a target element.

For example, the first data set includes information related to a morphological value (or morphological index, V1) calculated from a region corresponding to the first part of the first brain image IM1 acquired under the first scan condition and the first data set. Information related to a morphological value (or a morphological index, V2) calculated from a region corresponding to the second portion of the first brain image IM1 acquired under one scan condition may be included. In addition, the first data set includes information related to a morphological value (or morphological index, V1') calculated from a region corresponding to the first part of the second brain image IM2 acquired under the second scan condition and the second data set. Information related to a morphological value (or morphological index, V2') calculated from a region corresponding to the second portion of the second brain image IM2 acquired under the two-scan condition may be included.

In addition, the plurality of brain image data may further include various brain data similar to the first data set. For example, the n-th data set includes information related to a morphological value (or morphological index, V1n) calculated from a region corresponding to the first region of the n-th brain image IMn acquired under the first scan condition, and Information related to a morphological value (or a morphological index, V2n) calculated from a region corresponding to the second region of the first brain image IM1n acquired under the first scan condition may be included. In addition, the n-th data set includes information related to a morphological value (or morphological index, V1n') calculated from a region corresponding to the first region of the second brain image IM2n acquired under the second scan condition and the second It may include information related to a morphological value (or morphological index, V2n') calculated from a region corresponding to the second region of the second brain image IM2n acquired under the two-scan condition.

In this case, the second portion may be located adjacent to the skull compared to the first portion. That is, the first region and the second region may be brain elements present at different positions within the skull.

Also, the first region and the second region may be anatomically distinguished.

Alternatively, the first site and the second site may perform different functions. In other words, the first region may be a brain element performing a first brain function, while the second region may be a brain element performing a second brain function.

Referring back to FIG. 60 , the method of obtaining a correction parameter according to an embodiment of the present application includes the step of analyzing the correlation of morphological values according to the scan condition of at least one data set ( S6200 ) and the correlation analysis result. It may include obtaining a correction parameter based on the step (S6300).

In this case, the method of obtaining the correction parameter according to an embodiment of the present application may perform correlation analysis based on morphological values according to the scan condition of at least one data set and the “location where the target element is distributed”.

See FIG. 65 . 65 is an exemplary diagram of correlation analysis between morphological values according to the location of a target element according to an embodiment of the present application. In this case, the x-axis may be a morphological value obtained under the first scan condition, and the y-axis may be a morphological value obtained under the second scan condition.

As an example, referring to FIG. 65 , the correction parameter obtaining apparatus 2400 is configured under the first scan condition (eg, 1.5T) in relation to the first region (eg, frontal lobe region) included in the arbitrary i-th data set. ( A first correction parameter for the first region may be obtained based on the V1i').

As another example, referring to FIG. 65 , the correction parameter obtaining apparatus 2400 is configured under the first scan condition (eg, 1.5T) with respect to the second region (eg, cerebellum region) included in the arbitrary j-th data set. Morphological values (V2j) obtained from the j-th brain image (IM1j) taken at '), a second correction parameter for the second region may be obtained.

In this case, the first correction parameter for the first region and the second correction parameter for the second region may be different.

For example, the first calibration parameter may include a parameter that approximates a morphological value related to the first site obtained under the first scan condition to a morphological value related to the first site measured under the second scan condition. Alternatively, the first correction parameter may include a parameter that approximates a morphological value related to the first site obtained under the second scan condition to a morphological value related to the first site measured under the first scan condition.

For example, the first correction parameter is obtained from a transformation function obtained through correlation analysis between a morphological value related to the first site obtained under the first scan condition and a morphological value related to the first site obtained under the second scan condition. can be For example, if the linear function obtained through the correlation analysis between the first morphological value and the second morphological value is y=a1*x+b1, a1 and b1 are the first regions obtained under the first scan condition. It may be a parameter for approximating the morphological value of α to the morphological value of the first site obtained under the second scan condition.

On the other hand, the second correction parameter is determined through correlation analysis based on the morphological value related to the “second site” obtained under the first scan condition and the morphological value related to the “second site” obtained under the second scan condition, It may include a parameter for approximating the morphological value related to the second site obtained under the first scan condition to the morphological value related to the second site measured under the second scan condition. Alternatively, the second correction parameter may include a parameter that approximates the morphological value related to the second site obtained under the second scan condition to the morphological value related to the second site measured under the second scan condition.

For example, the second correction parameter is obtained from a transformation function obtained through correlation analysis between a morphological value related to a second site obtained under the first scan condition and a morphological value related to a second site obtained under the second scan condition. can be For example, if the linear function obtained through the correlation analysis between the first morphological value and the second morphological value is y=a2*x+b2, then a2 and b2 are the second regions obtained under the first scan condition. It may be a correction parameter for approximating the morphological value of α to the morphological value of the second region obtained under the second scan condition.

In addition, the correction parameter obtaining apparatus 2400, similarly as described above, through correlation analysis, obtains the morphological value of the first site obtained under the third scan condition from the morphological value of the first site obtained under the second scan condition. A correction parameter for approximating to , and a correction parameter for approximating the morphological value of the second region obtained under the second scan condition to the morphological value of the second region obtained under the third scan condition may be obtained.

In addition, the correction parameter obtaining apparatus 2400, similarly as described above, through a correlation analysis, the morphological value of the first site obtained under the first scan condition is the morphological value of the first site obtained under the third scan condition. A correction parameter for approximating to , and a correction parameter for approximating the morphological value of the second region obtained under the first scan condition to the morphological value of the second region obtained under the third scan condition may be obtained.

The correction parameter obtaining apparatus 2400 according to an embodiment of the present application may more accurately obtain the correction parameters for each part in consideration of the variable for the position of the target element, and a morphological value affected by the scan condition for each part. There is a beneficial effect of compensating the index.

Accordingly, the image analysis apparatus 2000 according to an embodiment of the present application has an advantage in that it can provide a more accurate and objective morphological index related to a brain disease and an auxiliary index for diagnosing a brain disease to the user.

Additionally, the correction parameter acquisition apparatus 2400 according to an embodiment of the present application performs the above-described correction parameter acquisition operation with respect to target elements (eg, the first brain element and the second brain element) existing in both the left and right hemispheres of the brain. Each can be implemented to perform.

For example, in an example related to FIG. 65 , the correction parameter obtaining apparatus 2400 may obtain the correction parameter by performing correlation analysis on the first region (eg, frontal lobe) located in the left hemisphere.

In addition, the correction parameter obtaining apparatus 2400 may be implemented to obtain the correction parameters by performing correlation analysis on the first region (eg, frontal lobe) located in the right hemisphere, and the correction obtained for the first region The parameters may be different depending on the left and right hemispheres.

On the other hand, the morphological value or morphological index related to the target element output from the image analysis apparatus 2000 is a morphological value or morphological index corresponding to a scan condition targeted by the user of the image analysis apparatus 2000 . may need to be modified. For example, when the output target element-related morphological value, etc. is a value obtained based on the morphological value obtained under the first scan condition, the user may use the morphology obtained under the second scan condition instead of the first scan condition. You may want to output morphological information related to a target element as results corresponding to values.

The image analysis apparatus 2000 according to an embodiment of the present application may obtain an input related to a reference scan condition from a user through an input module, and may be implemented to output morphological information related to a target element corresponding to the user input. can

Hereinafter, with reference to FIG. 66 , another embodiment of correcting a brain morphological index based on a user input related to a reference scan condition will be described in more detail. FIG. 66 is a diagram illustrating a flowchart of an image analysis method according to an embodiment of the present application, and in detail, is a diagram in which step S5300 of FIG. 58 is subdivided.

Referring to FIG. 66 , the step of correcting the brain morphological index according to an embodiment of the present application ( S5300 ) includes obtaining a target morphological value of the target element based on the segmentation result ( S5410 ), reference scan condition It may include the step of obtaining (S5420) and the step of determining whether the scan condition of the target image and the reference scan condition correspond (S5430).

At this time, if the scan condition of the target image and the reference scan condition correspond, the step of correcting the brain morphological index according to an embodiment of the present application ( S5300 ) is to output the brain morphological index based on the target morphological value. It may include step S5440.

On the other hand, if the scan condition of the target image and the reference scan condition do not correspond, correcting the brain morphology index according to an embodiment of the present application (S5300) includes obtaining a correction parameter (S5450), the target morphology The method may include obtaining a target morphological correction value based on the value and the correction parameter ( S5460 ) and outputting a brain morphological index based on the target morphological correction value ( S5470 ).

In operation S5410 of acquiring the target morphological value of the target element based on the segmentation result, the image analysis apparatus 2000 may perform the segmentation operation on the basis of voxel data of the target image related to the region corresponding to the target element. Thus, it is possible to obtain a target morphological value related to the target element.

Alternatively, the image analysis apparatus 2000 may acquire a morphological value related to the internal region of the skull based on voxel data of the target image related to the internal region of the skull obtained by the segmentation operation.

In addition, in the step of obtaining the target morphological value of the target element based on the segmentation result ( S5410 ), the image analysis apparatus 2000 , based on the morphological value related to the target element and the morphological value related to the inner region of the skull Thus, it is possible to obtain a brain morphological index of the target element related to the target element.

In this case, the target morphological value or the brain morphological index may be a numerical value related to a morphological character including volume, thickness, length, and shape.

In the step of obtaining the reference scan condition ( S5420 ), the image analysis apparatus 2000 may obtain an input related to the user's reference scan condition through the input module 2040 included in the image analysis apparatus 2000 . Alternatively, the image analysis apparatus 2000 may obtain an input related to the reference scan condition of the user through the input module 2640 of the output apparatus 2600 .

Alternatively, the reference scan condition may be preset. The image analysis apparatus 2000 may acquire information related to a preset reference scan condition.

In this case, the reference scan condition may be a scan condition related to a magnetic field strength, a setting parameter of the imaging device, or a manufacturer of the imaging device. Alternatively, the reference scan condition may be a scan condition that additionally considers a location of a target element for which a morphological index is to be obtained.

The image analysis apparatus 2000 may output or correct a morphological value or a brain morphological index of a target element corresponding to the reference scan condition by obtaining the reference scan condition. For example, by inputting the reference scan condition, the user may be provided with a morphological value or a brain morphological index associated with a target element estimated as a value measured under the reference scan condition.

In the step of determining whether the scan condition of the target image and the reference scan condition correspond ( S5430 ), the image analysis apparatus 2000 may determine whether the obtained reference scan condition and the scan condition in which the target image is captured correspond to each other have.

Specifically, the image analysis apparatus 2000 may obtain information related to a scan condition in which a target image is captured from the image acquisition apparatus 1000 or any external device (eg, a server).

For example, as described above, a scan condition in which a target image is captured may be structured as metadata with respect to the target image, and the image analysis apparatus 2000 recognizes metadata related to a scan condition in which the target image is captured. Information related to a scan condition in which the target image is captured may be acquired.

As another example, the image analysis apparatus 2000 may obtain a scan condition in which the target image is captured by receiving a scan condition in which the target image is captured from the user through the input module.

For another example, as described above, information related to a scan condition in which a target image is captured may be obtained from an arbitrary external device.

In this case, the image analysis apparatus 2000 may determine whether a scan condition in which the acquired target image is captured and a reference scan condition correspond.

For example, if the target image is photographed at a magnetic field strength of 1.5T and information related to a magnetic field strength 3T is obtained from the user as a reference scan condition, the image analysis apparatus 2000 corresponds to the scan condition of the target image and the reference scan condition It can be concluded that it does not

On the other hand, if the target image is photographed at a magnetic field strength of 3T and a reference scan condition related to the magnetic field strength 3T is obtained, the image analysis apparatus 2000 may determine that the scan condition of the target image and the reference scan condition correspond.

As another example, if the target image is photographed by an image device manufactured by a first manufacturer and a reference scan condition related to the second manufacturer is obtained, the image analysis apparatus 2000 determines that the scan condition of the target image and the reference scan condition are It can be judged that there is no correspondence.

On the other hand, if the target image was taken by the image device manufactured by the first manufacturer and the reference scan condition related to the first manufacturer is obtained, the image analysis apparatus 2000 determines that the scan condition of the target image and the reference scan condition correspond can do.

In the step of determining whether the scan condition of the target image and the reference scan condition correspond ( S5430 ), if the image analysis apparatus 2000 determines that the scan condition of the target image matches the reference scan condition, the image analysis apparatus 2000 may be implemented to output the brain morphological index based on the target morphological value of the target element obtained based on the segmentation result. (S5440)

In other words, if the scan condition of the target image and the reference scan condition are matched, the image analysis apparatus 2000 determines the target element obtained based on the segmentation result because correction to the scan condition desired by the user may not be performed. It may be implemented to output a brain morphological index based on a target morphological value.

On the other hand, in the step of determining whether the scan condition of the target image and the reference scan condition match ( S5430 ), if the image analysis apparatus 2000 determines that the scan condition of the target image and the reference scan condition do not correspond, the image analysis apparatus Reference numeral 2000 may be implemented to obtain a correction parameter obtained based on a correlation analysis corresponding to the scan condition of the target image and the reference scan condition from the correction parameter obtaining apparatus 2400 . (S5450)

For example, if the target image is photographed at a magnetic field strength of 1.5T, and a reference scan condition related to a magnetic field strength of 3T is obtained, the image analysis apparatus 2000 may obtain a magnetic field strength from the correction parameter obtaining apparatus 2400 at 1.5T. It may be implemented to obtain a correction parameter that converts or approximates a first morphological value obtained from the photographed first image to a second morphological value obtained from a second image photographed at 3T with a magnetic field strength. Alternatively, the image analysis apparatus 2000 may obtain a first morphological value obtained from a first image photographed at a magnetic field strength of 1.5T from a correction parameter database obtained from the correction parameter obtaining apparatus 2400 when the magnetic field strength is taken at 3T. A parameter that transforms or approximates the second morphological value obtained from the second image may be determined as a correction parameter.

As another example, if the target image is captured by an image device manufactured by a first manufacturer and a reference scan condition related to the second manufacturer is obtained, the image analysis apparatus 2000 may obtain a second image from the correction parameter obtaining apparatus 2400 . A correction parameter that converts or approximates a first morphological value obtained from a first image photographed with an imaging device of a 1 manufacturer and a second morphological value obtained from a second image photographed with an imaging device of a second manufacturer It can be implemented to obtain Alternatively, the image analysis apparatus 2000 may include a first morphological value obtained from a first image photographed with an image apparatus of a first manufacturer from a correction parameter database obtained from the apparatus for obtaining a correction parameter 2400 and an image apparatus of a second manufacturer. It may be implemented to determine, as a correction parameter, a parameter that is converted or approximated to a second morphological value obtained from the captured second image taken with .

In the step of obtaining the target morphological correction value based on the target morphological value and the correction parameter ( S5460 ), the image analysis apparatus 2000 obtains the target morphological value of the target element obtained based on the obtained correction parameter and the segmentation result. A target morphological correction value (or a corrected target morphological value) may be obtained based on .

For example, if the target image is photographed at a magnetic field strength of 1.5T and a reference scan condition related to a magnetic field strength of 3T is obtained, the image analysis apparatus 2000 has a shape related to the target element obtained from the brain image photographed at 1.5T A first correction parameter capable of correcting the morphological value (or morphological indicator) into a morphological value (or morphological indicator) that can correspond to that obtained in 3T may be acquired. In this case, the image analysis apparatus 2000 obtains the target morphological correction value by applying the first correction parameter to the target morphological value (or target morphological index) related to the target element obtained based on the segmentation result of the target image. can do.

As another example, if the target image is captured by an image device manufactured by a first manufacturer and a reference scan condition related to a second manufacturer is obtained, the image analysis device 2000 may display a brain photographed by an image device manufactured by the first manufacturer. A second correction parameter capable of correcting a morphological value (or morphological index) related to a target element obtained from an image into a morphological value (or morphological index) that can correspond to that obtained from an image device of a second manufacturer can be obtained. In this case, the image analysis apparatus 2000 obtains the target morphological correction value by applying the second correction parameter to the target morphological value (or target morphological index) related to the target element obtained based on the segmentation result of the target image. can do.

In step S5470 of outputting the brain morphological index based on the target morphological correction value, the image analysis apparatus 2000 may output the brain morphological index related to the target element based on the target morphological correction value.

For example, the image analysis apparatus 2000 calculates a brain morphological index related to a target element based on a morphological value corresponding to the internal region of the skull obtained according to the segmentation result of the target image and a target morphological correction value. It can be implemented to output. More specifically, the image analysis apparatus 2000 may be implemented to calculate and output the ratio of the target morphological correction value to the morphological value corresponding to the inner region of the skull as a brain morphological index related to the target element.

In this case, the morphological value corresponding to the inner region of the skull may be a value corrected by obtaining a correction parameter according to the user's input of the reference scan condition as described above.

As described above, as the scan conditions, the magnetic field strength and the content related to the manufacturer of the image device have been mainly described, but the present invention is not limited thereto, and scan conditions for setting parameters related to the image device may be similarly applied.

Also, according to FIG. 66 , the image analysis apparatus 2000 determines whether the scan condition of the target image and the reference scan condition correspond, and has been described as acquiring or determining the correction parameter according to the determination result, but is not limited thereto. , a process in which the image analysis apparatus 2000 determines whether the scan condition of the target image corresponds to the reference scan condition may be omitted. For example, the image analysis apparatus 2000 obtains a reference scan condition and a scan condition of the target image, and a target morphological correction value or a brain morphological index based on a correction parameter determined based on the reference scan condition and the scan condition of the target image It may be implemented to obtain

Hereinafter, an example of information related to a morphological index output and a user interface for receiving an input related to a reference scan condition of a user according to an embodiment of the present application will be described with reference to FIG. 67 . 67 is a diagram schematically illustrating a user interface according to an embodiment of the present application.

Referring to FIG. 67 , the user interface according to an embodiment of the present application obtains a user input related to a reference scan condition through an input module 2040 or 2640 , and receives a reference scan condition through an output module 2050 or 2650 . It may be provided to output information corresponding to a user input related to the . In this case, the output module 2050 or 2650 may be implemented in any suitable form, such as a display of a smartphone or a display of a monitor.

For example, the image analysis apparatus 2000 may output information related to the morphological index through the output module 2050 of the image analysis apparatus 2000 or the output module 2650 of the output apparatus 2600 .

In this case, the image analysis apparatus 2000 may be implemented to output information O1 about the current scan condition. In detail, the image analysis apparatus 2000 may be implemented to output information O1 related to which scan condition the outputted information related to the morphological index is outputted. For example, referring back to FIG. 67 , the information related to the currently output morphological index indicates that the target image was obtained under the scanning condition corresponding to the magnetic field strength of 1.5T in the imaging device manufactured by the first manufacturer. may contain information.

The user interface according to an embodiment of the present application may be implemented to output information O2 about the reference scan condition including checkboxes for obtaining an input for the user's input of the reference scan condition.

Specifically, a user interface may be implemented so that a user selects a checkbox corresponding to a reference scan condition related to magnetic field strength through an input module, and the image analysis apparatus 2000 obtains a reference scan condition related to magnetic field strength through the user interface can do.

In addition, a user interface may be implemented so that the user selects a check box corresponding to the reference scan condition related to the manufacturer through the input module, and the image analysis apparatus 2000 obtains the reference scan condition related to the manufacturer through the user interface. can The information O2 on the reference scan condition may be input through an input interface other than a check box.

For example, referring to FIG. 67 , a user may use information related to a morphological index currently output through an input module (eg, under a scan condition corresponding to a magnetic field strength of 1.5T in an image device manufactured by the first manufacturer) Input data for converting the calculated information ( F1 , G1 , T1 ) into information related to a morphological index as obtained under a scan condition corresponding to a magnetic field strength of 3T in an imaging device manufactured by the first manufacturer can be entered.

In this case, the image analysis apparatus 2000 obtains a corresponding correction parameter based on the obtained reference scan condition described above in relation to FIG. 67, and corrects and outputs the morphological value (or morphological index). It can be implemented to perform

In this case, the image analysis apparatus 2000 may be implemented to change and output previously output information related to the morphological index based on the corrected morphological value or the corrected morphological index.

For example, the calculated morphological value or morphological index may be changed according to the reference scan condition obtained from the user, and based on this, the brain volume percentile graph G1 is modified and outputted through the output module 2050 . can be implemented.

For another example, the calculated morphological value or morphological index may be changed according to the reference scan condition obtained from the user, and based on this, the morphological value or morphological index included in the detailed brain volume analysis table T1 and Numerical values such as related percentiles may be modified and implemented to be output through the output module 2050 .

As another example, the calculated morphological value or morphological index may be changed according to the reference scan condition obtained from the user, and based on this, the percentile associated with the morphological index included in the degree of brain volume atrophy (F1) is calculated. It may be implemented so that the corresponding color is corrected and outputted through the output module 2050 .

However, the user interface shown in FIG. 67 is only an example, and the user interface acquires any appropriate input from the user through any suitable method, and displays the analysis result of the image analysis device that is changed based on the user input in any suitable manner. It may be implemented to output in a method.

According to the present embodiment, the image analysis apparatus 2000 sets a target morphological value corresponding to a region related to a target element of a target image photographed under the first scan condition, based on a user input, other than the first scan condition. It can be calculated as a morphological estimate of the target element under the second scan condition or the third scan condition.

Accordingly, the image analysis apparatus 2000 according to the present embodiment can freely correct a target morphological value into a morphological estimate value corresponding to a scan condition desired by the user, and thus provides morphological index information according to the scan condition desired by the user to the user. There is an advantageous effect that it can provide.

Meanwhile, an example for correcting a morphological value according to an embodiment of the present application may be performed even before the segmentation step S5200 of FIG. 58 . In other words, in order to more accurately obtain a morphological value, the image analysis apparatus 2000 may be implemented to perform a correction operation before the segmentation step ( S5200 ).

In this case, according to an embodiment of the present application, correction related to the target image may be implemented differently for each scan condition in which the target image is obtained.

As an example, the image analysis apparatus 2000 may be provided to perform an operation such as regularization of the intensity of the target image. In this case, a correction method of regularization of intensity may be implemented differently for each scan condition in which the target image is obtained.

As another example, the image analysis apparatus 2000 may be implemented to convert a target image into a brain image corresponding to a scan condition related to a target magnetic field strength or a setting parameter. For example, the image analysis apparatus 2000 converts a target image into a brain image corresponding to a target scan condition using any suitable software such as an MR simulator, an image conversion technique, or a learned artificial neural network. can do. At this time, the target image is corrected by providing different image conversion parameters such as the MR simulator for each scan condition in which the target image was acquired, particularly in consideration of the setting parameters (eg, TR, TE, etc.) in which the target image was acquired. This can be done. In this case, the image analysis apparatus 2000 may be implemented to obtain or calculate a conversion parameter for converting the target image into a brain image corresponding to the target scan condition in consideration of the scan condition.

Alternatively, the correction related to the target image may be differently provided for each position of the target element.

For example, the image analysis apparatus 2000 performs pre-processing on a first region corresponding to a first brain element by a first method, but with respect to a second region corresponding to a second brain element, a second region corresponding to a second brain element is performed in a second method different from the first method. It can be implemented to perform the preprocessing of the two methods.

For example, the first region corresponding to the first brain element may be located adjacent to the skull region compared to the second region corresponding to the second brain element. In this case, in relation to the resolution or sharpness of the target image, the first region may not be relatively clear compared to the second region. In this case, the image analysis apparatus 2000 may be implemented to perform pre-processing for improving the sharpness of the first region by using the first method for the first region, and any suitable second method for the second region may be implemented. It can be implemented to perform pre-processing to improve the accuracy of image analysis using

As described above, the image analysis apparatus 2000 according to an embodiment of the present application not only corrects morphological values (eg, morphological values or morphological indicators) obtained based on the segmentation result, but also performs segmentation prior to segmentation. Since additional correction may be performed in consideration of scan conditions, etc., there may be an advantageous effect of obtaining more accurate morphological values.

Accordingly, the image analysis apparatus 2000 according to an embodiment of the present application may have an advantage in that it can provide a user with an objective diagnostic aid index related to a brain disease.

Various index information may be obtained through analysis of the medical image, and medical information that may be derived from the index information may also be various. For example, the medical information that can be derived from the indicator information may relate to diagnostic information, analysis information, or prescription information obtained through analysis of the indicator information.

In this case, information required for a user among various pieces of acquired indicator information may be different for each user. Accordingly, there is a need to selectively provide index information necessary for a user among various index information included in a medical image. Alternatively, there is also a need to selectively provide index information selected according to a user's interest among various index information included in a medical image. That is, by providing the necessary information to the user differently according to patient information or user information, as compared to uniformly providing index information derived from a medical image, the user's information acquisition may be easier.

According to an exemplary embodiment, the image output apparatus may selectively provide the user with the required index information from among various index information obtained through analysis of the photographed medical image.

68 is a diagram for describing a medical information output process according to an embodiment.

Referring to FIG. 68 , the medical information output process according to an embodiment includes a medical data acquisition step (S4100), a medical information acquisition step (S4200), a diagnosis auxiliary information acquisition step (S4300), and a diagnosis auxiliary information output step (S4400). may include

69 is a diagram for describing an image analysis apparatus according to an exemplary embodiment.

Referring to FIG. 69 , the image analysis apparatus 4000 according to an embodiment may output medical information. The image analysis apparatus 4000 may include a medical data acquisition module 4100 , a medical information acquisition module 4200 , a diagnosis auxiliary information acquisition module 4300 , and a diagnosis auxiliary information output module 4400 .

Although the drawings indicate that the medical data is acquired through the image analysis device, the medical data may also be acquired by the image acquisition device.

70 is a diagram for explaining a medical data acquisition module.

Referring to FIG. 70 , the medical data acquisition module 4100 may include either an image data acquisition unit 4110 or a non-image data acquisition unit 4130 . Medical data may include either image data or non-image data.

The image data acquisition unit 4110 may acquire image data. Here, the image data may include various types of medical images or medical images. The image data may include a plurality of medical images or medical images.

The non-image data acquisition unit 4130 may acquire non-image data. Here, the non-image data may include health-related questionnaire data. The non-image data may include non-image information obtained based on the image data.

71 is a diagram for explaining the medical information acquisition module 4200 .

The medical information acquisition module 4200 may acquire medical information based on medical data. The medical information may include various information related to a health state of the subject. The medical information may include various index information related to the object, for example, an area, volume, location, or shape of an area of the human body.

Referring to FIG. 71 , the medical information obtaining module 4200 may include either a first medical information obtaining unit 4210 or a second medical information obtaining unit 4230 .

The first medical information acquisition unit 4210 may acquire first medical information. The second medical information acquisition unit 4230 may acquire second medical information. In this case, the first medical information or the second medical information may include information obtained based on medical data using a learned neural network. The first medical information or the second medical information may include medical data and raw data. The first medical information or the second medical information may include a morphological index or a morphological value obtained based on the medical data.

The first medical information acquisition unit 4210 and the second medical information acquisition unit 4230 may acquire medical information based on independent medical data, respectively. The first medical information obtaining unit 4210 may obtain medical information based on the first medical data, and the second medical information obtaining unit 4230 may obtain medical information based on the second medical data. For example, the first medical information obtaining unit 4210 may obtain medical information based on image data, and the second medical information obtaining unit 4230 may obtain medical information based on non-image data.

Alternatively, the first medical information acquisition unit 4210 and the second medical information acquisition unit 4230 may acquire medical information based on common medical data. The first medical information acquisition unit 4210 and the second medical information acquisition unit 4230 may acquire medical information based on the first medical data. For example, the first medical information acquisition unit 4210 and the second medical information acquisition unit 4230 may acquire medical information based on image data or non-image data.

In FIG. 71 , the first medical information obtaining unit 4210 and the second medical information obtaining unit 4230 are exemplified as being distinguished, but this is only an example and the first medical information obtaining unit 4210 and the second medical information obtaining unit 4210 are exemplified. The unit 4230 may be provided as one physical or logical configuration.

72 is a diagram for explaining the diagnostic auxiliary information obtaining module 4300 .

The diagnostic auxiliary information obtaining module 4300 may obtain diagnostic auxiliary information. The diagnosis auxiliary information may be information obtained based on medical information using a learned neural network. The diagnosis auxiliary information may be information obtained by processing medical information. For example, the diagnosis auxiliary information may include disease diagnosis information based on medical information, medical information analysis information, prescription information, and the like. The diagnostic auxiliary information may be in the form of an image or text.

In addition, the diagnosis auxiliary information may include information related to image analysis exemplified herein, for example, image quality related information, ICV related information, calibration related information, or segmentation related information.

Referring to FIG. 72 , the auxiliary diagnosis information obtaining module 4300 may include either the first diagnosis auxiliary information obtaining unit 4310 or the second diagnosis auxiliary information obtaining unit 4330 .

The first auxiliary diagnosis information obtaining unit 4310 may obtain first auxiliary diagnosis information. The second diagnosis auxiliary information obtaining unit 4330 may obtain second diagnosis auxiliary information.

The first diagnosis auxiliary information obtaining unit 4310 and the second diagnosis auxiliary information obtaining unit 4330 may respectively obtain diagnosis auxiliary information based on independent medical information. The first auxiliary diagnosis information obtaining unit 4310 may obtain auxiliary diagnosis information based on the first medical information, and the second auxiliary diagnosis information obtaining unit 4330 may obtain auxiliary diagnosis information based on the second medical information. have. Exemplarily, the first auxiliary diagnosis information obtaining unit 4310 obtains diagnosis auxiliary information based on image data, and the second auxiliary diagnosis information obtaining unit 7313 obtains diagnosis auxiliary information based on non-image data. can do.

Alternatively, the first diagnosis auxiliary information obtaining unit 4310 and the second diagnosis auxiliary information obtaining unit 4330 may obtain the diagnosis auxiliary information based on common medical information. The first diagnosis auxiliary information obtaining unit 4310 and the second diagnosis auxiliary information obtaining unit 4330 may obtain diagnosis auxiliary information based on the first medical information. Exemplarily, the first medical information may be image data or non-image data.

In FIG. 72 , the first diagnosis auxiliary information obtaining unit 4310 and the second diagnosis auxiliary information obtaining unit 4330 are exemplified as being distinguished, but this is only an example and the first diagnosis auxiliary information obtaining unit 4310 and the second diagnosis auxiliary information obtaining unit 4310 are distinguished. The diagnostic auxiliary information obtaining unit 4330 may be provided in one physical or logical configuration.

73 is a diagram for explaining the diagnostic auxiliary information output module 4400 .

Referring to FIG. 73 , the auxiliary diagnosis information output module 4400 may include a diagnosis auxiliary information output unit 4410 , a comment generator 4420 , or a comment output unit 4430 .

The diagnostic auxiliary information output unit 4410 may output diagnostic auxiliary information. The diagnosis auxiliary information output unit 4410 may output a plurality of pieces of auxiliary diagnosis information (eg, prescription information and target disease related indicator information). For example, the diagnosis auxiliary information output unit 4410 may output information based on the first diagnosis auxiliary information and the second diagnosis auxiliary information. The diagnosis auxiliary information output unit 4410 may output information obtained by processing the diagnosis auxiliary information into an image or non-image form.

The comment generator 4420 may generate a comment determined based on the diagnosis auxiliary information. The comment generator 4420 may generate a determined comment based on a plurality of pieces of diagnostic auxiliary information. For example, the comment generator 4420 may generate a first comment based on the first auxiliary diagnosis information related to the first disease and the second diagnosis auxiliary information related to the second disease. In this case, the first comment may include information related to the first disease and/or information related to a prescription or medical treatment for the first disease.

The first comment may be generated based on the first auxiliary information related to dementia, for example, information about the degree of contraction of the hippocampus, and the second auxiliary information related to dementia, for example, information related to the shape of the hippocampus.

The first comment may indicate a message that the subject may correspond to an early stage of dementia with a dementia risk of 00%. Also, the first comment may indicate a message that cognitive ability may be reduced because the subject is in an early stage of dementia.

Meanwhile, the diagnosis auxiliary information output module may differently output diagnosis auxiliary information or comments according to a condition satisfied by the acquired medical information based on a plurality of predetermined conditions.

For example, the plurality of predetermined conditions include a first condition including a reference value for the one or more medical information and a second condition including a reference value for the one or more medical information and at least partially different from the first condition, and the diagnostic auxiliary information The output module outputs first diagnosis auxiliary information (or a first comment) when the plurality of pieces of acquired medical information satisfies a first condition, and a second diagnosis when the plurality of pieces of obtained medical information satisfies the second condition Supplementary information (or second comment) may be output.

The first condition is related to the first disease, and may include a first condition for the first medical information and a second condition for the second medical information. The first condition for the first medical information may include that the first medical information exceeds the first reference value, and the second condition for the second medical information may include that the second medical information exceeds the second reference value. have. The second condition is related to the second disease and may be implemented similarly to the first condition.

Hereinafter, predetermined conditions that are based on generation of diagnostic auxiliary information and/or comments will be described with reference to some embodiments.

74 is a diagram for illustrating related indicators for some target diseases.

Referring to FIG. 74 , in the case of Alzheimer's dementia, it may be related to a temporal lobe atrophy index, a parietal atrophy index, and a hippocampal atrophy index. In the case of vascular dementia, it may be associated with an indicator of temporal lobe atrophy, an indicator of parietal atrophy, an indicator of hippocampal atrophy, an indicator of white matter lesion atrophy, and an indicator of gyrus atrophy. In the case of frontotemporal lobe (FTLD) dementia, it may be associated with indicators of frontal lobe atrophy, indicators of temporal lobe atrophy, and indicators of hippocampal atrophy.

Correlation between each disease and indicator may be different. For example, the hippocampal atrophy index may have a higher correlation with Alzheimer's dementia than the temporal and parietal atrophy index. The white matter lesion index may have a higher correlation with vascular dementia than the hippocampal atrophy index, and the hippocampal atrophy index may have a higher correlation with vascular dementia than the frontal lobe atrophy index, the temporal lobe atrophy index, and the parietal lobe atrophy index. The frontal atrophy index and the temporal atrophy index may have a higher correlation with frontotemporal dementia than the hippocampal atrophy index.

The diagnosis auxiliary information output module 4400 may output the diagnosis auxiliary information generated based on the correlation between an index (medical information) and a disease as illustrated in FIG. 74 .

For example, referring to FIG. 74 , when the frontal lobe index, the temporal lobe index, and the hippocampal index are each greater than or equal to a threshold value according to the medical information, the diagnostic auxiliary information output module 4400 provides diagnostic auxiliary information indicating that the subject belongs to a risk group for FTLD. can be printed out. Alternatively, when each of the hippocampal index, the temporal lobe index, and the parietal lobe index is equal to or greater than a threshold value, the diagnostic auxiliary information output module 4400 may output diagnostic auxiliary information indicating that the subject belongs to a risk group for Alzheimer's disease (AD).

75 is a diagram for explaining a medical information output screen according to an exemplary embodiment.

Referring to FIG. 75 , the medical information output screen may include patient information (PI), image information (RI), indicator information (FI), diagnosis information (DI), or comments (CMT).

The image information RI may include an image on which a medical image related to diagnosis auxiliary information is displayed. The indicator information FI may include information related to medical information obtained from medical data. The diagnosis information DI may include diagnosis auxiliary information obtained from medical information, for example, disease diagnosis information, medical information analysis information, prescription information, and the like.

The comment CMT may include a comment generated based on the diagnosis auxiliary information. Referring to FIG. 74 together, when the frontal lobe index, the temporal lobe index, and the hippocampal index are each greater than or equal to the threshold, 'Frontal- It is possible to provide a comment indicating that it is a situation in which the onset of the disease (temporal lobe dementia) is suspected.

Alternatively, when the hippocampal index, the temporal lobe index, and the parietal lobe index are above the threshold values, 'the degree of hippocampal atrophy is higher than that of normal individuals, and the degree of atrophy of the temporal and parietal lobes corresponds to a risk level, so AD (Alzheimer's disease) You can provide a comment indicating that the disease is 'suspicious'.

The comment CMT may include prescription information generated based on the diagnosis auxiliary information. 74 together, when the temporal lobe index, the parietal lobe index, and the hippocampal index are each greater than or equal to the threshold, 'Since Alzheimer's is suspected, the first prescription information, for example, a platelet aggregation inhibitor such as aspirin, an anticoagulant such as warfarin, blood flow It is possible to provide a comment indicating that the prescription of a circulation improving agent or an acetylcholinease inhibitor is required.

Alternatively, if the frontal index, the temporal index, and the hippocampal index are each greater than or equal to the threshold value, 'Since frontotemporal dementia is suspected, a second prescription, such as an antipsychotic drug, an antidepressant, an antianxiety agent, A comment indicating 'required' may be provided.

Various index information can be obtained through the analysis of medical images. When the various index information is uniformly displayed or provided to the user, the user (eg, a clinician) must selectively acquire the information as needed. discomfort exists. Accordingly, in order to improve user convenience, there is a need to provide different medical information according to user information or patient information.

The imaging device according to an embodiment may obtain medical data, acquire a plurality of medical information (eg, ICV or other morphological indices) based on the medical data, reconstruct the medical information, and then provide the reconstructed information. have.

The reconstructed information may be information reconstructed based on user information or patient information. Here, the user information includes a user's selection of medical subjects, categories, and preferred indicators, and the patient information includes patient personal information, information about the patient's health status, information obtained by analyzing the patient's medical data, The patient's past medical history may be included.

Alternatively, the reconstructed information may include information in which a priority or an arrangement order regarding a plurality of pieces of medical information obtained based on the above-described user information or patient information is determined.

Alternatively, the reconstructed information may include some medical information preferentially selected based on the above-described user information or patient information among a plurality of acquired medical information.

The provision of the reconstructed information by the diagnostic auxiliary information output module 4400 may include providing high-priority medical information to be displayed at the top based on user information or patient information. Alternatively, the provision of the reconstructed information by the diagnostic auxiliary information output module 4400 may include providing medical information selected based on the above-described user information or patient information.

Hereinafter, an apparatus and method for selectively providing medical information will be described in more detail.

76 is a diagram for explaining a medical information output process according to another embodiment.

Referring to FIG. 76, the medical information output process according to another embodiment may include a medical data acquisition step (S4100), a medical information acquisition step (S4200), a screening information acquisition step (S4700), and a screening information output step (S4900). can

77 is a diagram for explaining an image analysis apparatus according to another embodiment.

Referring to FIG. 77 , the image analysis apparatus 4000 according to another exemplary embodiment may output medical information. The image analysis apparatus 4000 may include a medical data acquisition module 4100 , a medical information acquisition module 4200 , a selection information acquisition module 4700 , and a selection information output module 4900 . Although the drawings indicate that the medical data is acquired through the image analysis device, the medical data may also be acquired by the image acquisition device.

The selection information acquisition module 4700 may acquire selection information.

The selection information may be information obtained based on medical information. The selection information may be information obtained based on medical information using a learned neural network model. The selection information may be information that is a basis for determining a disease among medical information. The selection information may refer to information selected using a neural network model learned from among medical information.

The selection information may refer to information selected by a predetermined selection criterion among medical information.

According to an embodiment, the selection criterion may be determined based on patient information. The selection criterion may be determined to select information related to a patient's condition from among medical information acquired based on image data, from medical information acquired based on image data or non-image data.

The selection criterion may be determined such that information related to a patient's condition is selected from among the medical information obtained based on the image data.

The selection criterion may be determined such that a predetermined medical information set is selected according to the reference medical information. As a specific example, when the reference medical information is the degree of contraction of the hippocampus, the selection criterion may be determined to output one or more predetermined medical information when the degree of contraction of the hippocampus is equal to or greater than the reference value. In this case, the selected one or more medical information may be medical information related to the contraction of the hippocampus. For example, the contraction of the hippocampus has a high correlation with vascular dementia, and the selected one or more medical information may include information on temporal lobe atrophy and parietal lobe atrophy information used for diagnosis of vascular dementia.

The selection criterion may be determined such that a predetermined medical information set is selected according to a user's selection. As a specific example, when the user belongs to the first department (eg, department of radiology), the selection criteria may be determined such that, when the user is the first department (eg, department of radiology), one or more predetermined medical information is output. . The selection information output module 4900 may selectively provide medical information corresponding to the selected department based on a user input for selecting one of a plurality of departments (or items). In this case, the selected one or more medical information may be medical information handled by the first medical department (eg, radiology department). For example, when the user belongs to the first department, the selection information output module 4900 may output medical information including the first indicator in response to a user input for selecting the first department. The first indicator may be predetermined to correspond to the first clinical department. The first department (eg, radiology department) may have a high correlation with the first indicator (eg, white matter lesion, etc.).

The selection criterion may be determined such that a predetermined set of medical information is selected according to the disease of interest of the doctor. As a specific example, when the doctor is interested in a first disease (eg, dementia), the selection criteria may be determined such that, when the doctor is the first disease (eg, dementia), one or more predetermined medical information is output. . The selection information output module 4900 may selectively provide medical information corresponding to the selected disease based on a user input for selecting any one of a plurality of diseases of interest. In this case, the selected one or more pieces of medical information may be medical information having a high relevance to the first disease (eg, dementia).

78 is a diagram for explaining a selection information obtaining module.

Referring to FIG. 78 , the selection information obtaining module 4700 may include a first selection information obtaining unit 4710 or a second selection information obtaining unit 4730 .

The first selection information acquisition unit 4710 may acquire first selection information selected according to a first selection criterion among medical information.

The second selection information acquisition unit 4730 may acquire second selection information selected according to the second selection criterion among the medical information.

The first selection criterion may be any one of criteria related to user information and patient information. The second selection criterion may be any one different from the first selection criterion among criteria related to user information or patient information. For example, the first selection criterion may be a criterion based on user information (eg, the type of the doctor's department, the doctor's interest), and the second selection criterion is the patient information (eg, patient personal information, the patient's medical data is analyzed) It may be a criterion based on information obtained through

79 is a diagram for explaining the selection information output module 4900 .

Referring to FIG. 79 , the selection information output module 4900 may include a selection information arranging unit 4910 , a selection information output unit 4920 , a comment generator 4930 , and a comment output unit 4940 .

The selection information aligning unit 4910 may sort the selection information obtained through the selection information obtaining module 4700 according to a predetermined sorting criterion. The sorting criterion may include a criterion determined so that the selection information is sorted in an order that most matches the above-described selection criterion.

The selection information aligning unit 4910 may sort the selection information according to a predetermined criterion so that the selection information is sorted by type of selection information. For example, the selection information aligning unit 4910 may include a criterion determined to sort and sort the selection information related to the first disease and the selection information related to the second disease. The selection information aligning unit 4910 may be omitted.

The selection information output unit 4920 may output selection information. The screening information output unit 4920 may output some of a plurality of pieces of medical information included in the screening information. The screening information output unit 4920 may output some information having a high degree of meeting the screening criteria among a plurality of pieces of medical information included in the screening information.

The comment generator 4930 may generate a comment based on the selection information. The comment generator 4930 may generate a comment based on a plurality of pieces of selection information. For example, the comment generator 4930 may generate a first comment based on the first selection information related to the first disease and the second selection information related to the first disease. In this case, the first comment may include information related to the first disease and/or information related to a prescription or medical treatment for the first disease.

For example, the first comment may be generated based on the first selection information related to dementia, eg, information about the degree of contraction of the hippocampus, and the second selection information related to dementia, eg, information related to the shape of the hippocampus. In this case, the first comment may indicate a message that the subject has a dementia risk of 00%, which may correspond to an early stage of dementia. Also, the first comment may indicate a message that cognitive ability may be reduced because the subject is in an early stage of dementia. Also, the first comment may indicate prescription information based on the subject's dementia progression level.

80 and 81 are diagrams for explaining a selection information output screen.

Referring to FIG. 80 , the selection information output screen according to an embodiment may be an output screen based on information selected according to a selection criterion determined to select information related to patient information from among medical information.

The screen for outputting screening information according to an embodiment may include patient information (PI), image information (MI), disease information (DI), screening information (SI), analysis information (AN), or comments (CMT).

The image information MI may include a medical image related to selection information selected according to a selection criterion (eg, a selection criterion according to patient information).

The disease information DI may include disease information acquired based on selection information selected according to the first selection criterion. In this case, the disease information DI may include information on at least one or more diseases that may occur in the object and information on a probability that the corresponding disease will occur in the object. The disease information DI may include information in which a plurality of disease information that may be generated in an object is arranged according to a predetermined criterion.

The selection information SI may include an anatomical index related to patient information among anatomical indices included in the medical information. The screening information SI may include an anatomical index related to the disease information DI. For example, when the user selects the first disease information from among the plurality of disease information, the selection information output screen may provide the first selection information regarding the anatomical index related to the first disease. When the user selects the first disease information and the second disease information from among the plurality of disease information, the selection information output screen may provide the first disease information and the second selection information on an anatomical index related to the second disease information. can

The analysis information AN may include a comprehensive analysis result obtained based on medical data about an object, for example, a comprehensive analysis result in a report format.

The comment CMT may include a comment generated based on selection information. Regarding comments, the above-described contents in this specification may be applied. The comment may generate a comment related to the first disease based on selection information obtained based on the first disease among the medical information. For example, referring to FIG. 80 , the comment may include first screening information (eg, degree of hippocampal atrophy) related to Alzheimer's disease, second screening information (eg, degree of frontal lobe atrophy) and information related to the subject's degree of Alzheimer's risk. have.

The comment may be changed in response to a user input for selecting any one of a plurality of disease information. For example, in FIG. 80 , as the user selects the vascular dementia disease information object, the comment may be changed to include selection information related to the subject's vascular dementia risk level and the subject's vascular dementia.

Referring to FIG. 81 , the screen for outputting screening information according to another exemplary embodiment may be an output screen based on information selected according to a screening criterion determined to select information related to a department from among medical information.

The screening information output screen according to another embodiment may include patient information (PI), image information (MI), department information (CAT), screening information (SI), analysis information (AN), or comments (CMT).

The image information MI may include a medical image related to selection information selected according to a selection criterion (eg, a selection criterion according to a medical subject).

The department information CAT may include disease information acquired based on selection information selected according to the second selection criterion. In this case, the department information CAT may include a department list that analyzes medical data related to the object. The department list may include at least one department requiring analysis of medical data related to the subject.

The screening information SI may include an anatomical index related to the selected department information CAT among anatomical indices included in the medical information. The screening information SI may include an anatomical index related to the selected clinical department information CAT). For example, when the user selects a first department from among a plurality of departments, the screen for outputting selection information may provide third selection information regarding anatomical indicators related to the first department. When the user selects the first department and the second department from among the plurality of departments, the screen for outputting screening information may provide fourth selection information regarding anatomical indicators related to the first department and the second department.

The analysis information AN may include a comprehensive analysis result obtained based on medical data about an object, for example, a comprehensive analysis result in a report format.

The comment CMT may include a comment generated based on selection information. Regarding the comments, the above-described contents in this specification may be applied. The comment may generate a comment related to the first department based on the selection information obtained in relation to the first department. For example, referring to FIG. 81 , the comment may include information related to first selection information and second selection information related to neurology.

The comment may be changed in response to a user's input for selecting any one of a plurality of medical departments. For example, in FIG. 81 , as the user selects an object of the department of radiology, the comment may be changed to include selection information required by the department of imaging.

The image analysis methods of the image analysis apparatus 2000 described above may be stored in the first memory 2020 of the image analysis apparatus 2000 , and the first controller 2030 of the image analysis apparatus 2000 may be stored in the first memory ( 2020) may be provided to perform image analysis methods stored in .

The brain image analysis method, brain image analysis apparatus, and brain image analysis system disclosed in the present application may be used to analyze a brain image.

In particular, the brain image analysis method, the brain image analysis apparatus, and the brain image analysis system disclosed in the present application can be applied to all fields that provide information related to brain diseases with high accuracy and reliability. For example, it may be used in a field of health check-up that calculates an auxiliary index for diagnosing a brain disease or provides an auxiliary index related to a brain disease.

However, the brain image analysis method, brain image analysis apparatus, and brain image analysis system disclosed in the present application may be applied not only to brain images but also to all medical images. Therefore, it can be applied to all fields for obtaining and providing morphological indicators related to the length, shape, volume, thickness, etc. of a specific region to assist in diagnosing various diseases as well as an auxiliary indicator for diagnosing brain diseases.

Features, structures, effects, etc. described in the above embodiments are included in at least one embodiment of the present invention, and are not necessarily limited to only one embodiment. Furthermore, features, structures, effects, etc. illustrated in each embodiment can be combined or modified with respect to other embodiments by those of ordinary skill in the art to which the embodiments belong. Accordingly, the contents related to such combinations and modifications should be interpreted as being included in the scope of the present invention.

In addition, although the embodiment has been mainly described in the above, this is only an example and does not limit the present invention, and those of ordinary skill in the art to which the present invention pertains to the above in the range that does not depart from the essential characteristics of the present embodiment. It will be appreciated that various modifications and applications not illustrated are possible. That is, each component specifically shown in the embodiment can be modified and implemented. And differences related to such modifications and applications should be construed as being included in the scope of the present invention defined in the appended claims.

Claims (22)

A device for acquiring a brain image and performing a morphological analysis based on the brain image is a method for analyzing a brain image, the method comprising:
Obtaining a first correction parameter for correcting a first morphological value related to a first brain element - The first correction parameter is a first correction parameter for correcting the first morphological value in consideration of a scan condition in which a brain image is taken contains 1 parameter-;
Obtaining a second correction parameter for correcting a second morphological value related to a second brain element different from the first brain element - The second correction parameter is the second morphological value according to a scan condition in which a brain image is taken including a second parameter to be corrected in consideration of -;
acquiring a target brain image;
The first region by segmenting the target brain image into a plurality of brain regions including a first region corresponding to the first brain element, a second region corresponding to the second brain element, and a skull region; obtaining a third region associated with the second region and an inner region of the cranial region;
acquiring a first brain morphological index related to the first brain element based on voxel data corresponding to the first region, voxel data corresponding to the third region, and the first correction parameter; and
acquiring a second brain morphological index related to the second brain element based on voxel data corresponding to the second region, voxel data corresponding to the third region, and the second correction parameter;
wherein the first morphological value, the second morphological value, the first brain morphological indicator, and the second brain morphological indicator are at least one of a volume, a length, a width, and a thickness of a brain element;
Brain image analysis method.
According to claim 1,
Obtaining the first brain morphological index or obtaining the second brain morphological index comprises:
The method further includes: obtaining a reference morphological value related to the inner region based on voxel data corresponding to the third region;
Acquiring the first brain morphological index comprises:
acquiring a first target morphological value related to the first brain element based on voxel data corresponding to the first region;
calculating a first corrected morphological value associated with a first brain element based on the first target morphological value and the first correction parameter; and
Calculating the first brain morphological index based on the first corrected morphological value and the reference morphological value;
Acquiring the second brain morphological index comprises:
acquiring a second target morphological value related to the second brain element based on voxel data corresponding to the second region;
calculating a second corrected morphological value associated with a second brain element based on the second target morphological value and the second correction parameter; and
Calculating the second brain morphological index based on the second corrected morphological value and the reference morphological value;
Brain image analysis method.
3. The method of claim 2,
The first morphological value and the second morphological value are values obtained under the first scan condition,
The target brain image is obtained under the first scan condition,
the first calibration parameter is a first parameter for calculating a third morphological value associated with the first brain element obtained under a second scan condition different from the first scan condition based on the first morphological value including,
the second calibration parameter comprises a second parameter for calculating a fourth morphological value associated with the second brain element obtained under the second scan condition based on the second morphological value;
The calculating the first corrected morphological value comprises: the first correction being a morphological estimate of the first brain element under the second scan condition, based on the first target morphological value and the first correction parameter. Including obtaining a morphological value,
Calculating the second corrected morphological value includes the second correction being a morphological estimate of the second brain element under the second scan condition based on the second target morphological value and the second correction parameter. Including obtaining a morphological value,
Brain image analysis method.
4. The method of claim 3,
the first calibration parameter comprises a first parameter set associated with a first linear function for calculating the third morphological value based on the first morphological value;
the second calibration parameter comprises a second parameter set associated with a second linear function for calculating the fourth morphological value based on the second morphological value;
wherein calculating the first corrected morphological value comprises obtaining the first corrected morphological value based on the first target morphological value and a first parameter set associated with the first linear function,
wherein calculating the second corrected morphological value comprises obtaining the second corrected morphological value based on the second target morphological value and a second set of parameters associated with the second linear function.
Brain image analysis method.
According to claim 1,
The first region corresponding to the first brain element is located adjacent to the skull region compared to the second region corresponding to the second brain element,
Brain image analysis method.
According to claim 1,
wherein the first brain component is associated with a component that performs a first brain function, and the second brain component is associated with a component that performs a second brain function that is different from the first brain function.
Brain image analysis method.
4. The method of claim 3,
The first scan condition and the second scan condition are a magnetic field strength related to the resolution of a brain image of the brain image acquisition device, a manufacturer of the brain image acquisition device, and a setting parameter related to a shape of a magnetic field generated by the brain imaging device associated with at least one of
Brain image analysis method.
According to claim 1,
The target brain image is obtained from a first object having a first characteristic,
The first correction parameter and the second correction parameter are obtained from a first brain image and a second brain image obtained from a second object having the first characteristic,
wherein the first characteristic is related to the age or sex of the subject;
Brain image analysis method.
According to claim 1,
The segmentation is performed using a neural network provided to obtain a plurality of brain regions corresponding to a plurality of brain elements including the first brain element, the second brain element, and the skull based on the target brain image. felled,
Brain image analysis method.
5. The method of claim 4,
Acquiring the target brain image is,
obtaining initial brain images;
Pre-processing the initial brain image; further comprising,
Pre-processing the initial brain image,
performing pre-processing on the first region of the initial brain image by a first method, and performing pre-processing of a second method different from the first method on the second region of the initial brain image,
The segmentation is performed based on the preprocessed target brain image,
Brain image analysis method.
5. The method of claim 4,
The first corrected morphological value is a volume value associated with the first brain element obtained with the target brain image, and the second target morphological value is a volume value associated with the second brain element obtained with the target brain image. is the value,
The first brain morphological index is calculated as the first corrected morphological value with respect to the reference morphological value,
wherein the second brain morphological index is calculated as the second corrected morphological value with respect to the reference morphological value.
Brain image analysis method.
A device for analyzing brain images, comprising:
an image acquisition unit for acquiring a target brain image; and
A controller that provides brain image analysis information based on the target brain image; including,
The controller is
A first correction parameter for correcting a first morphological value related to a first brain element - The first correction parameter is a first parameter for correcting the first morphological value in consideration of a scan condition in which a brain image is taken. a second correction parameter for obtaining a including- and correcting a second morphological value associated with a second brain element different from the first brain element, wherein the second correction parameter includes: Obtaining - including a second parameter to be corrected in consideration of the photographed scan condition, obtaining a target brain image, and applying the target brain image to a first region corresponding to the first brain element and the second brain element By performing segmentation into a plurality of brain regions including a corresponding second region and a cranial region, a third region associated with the first region, the second region, and an inner region of the cranial region is obtained, and the first region a first brain morphological index related to the first brain element is obtained based on voxel data corresponding to the region, voxel data corresponding to the third region, and the first correction parameter, and voxels corresponding to the second region configured to obtain a second brain morphological index associated with the second brain element based on data, voxel data corresponding to the third region, and the second correction parameter,
wherein the first morphological value, the second morphology, the first brain morphological indicator and the second brain morphological indicator are related to at least one of a volume, a length, a width, and a thickness of a brain element;
Brain image analysis device.
13. The method of claim 12,
The controller is
and to obtain a reference morphological value related to the inner region based on voxel data corresponding to the third region;
A first target morphological value related to the first brain element is obtained based on voxel data corresponding to the first region, and a first brain element and a first correction parameter are obtained based on the first target morphological value and the first correction parameter. calculating a related first corrected morphological value, calculating the first brain morphological index based on the first corrected morphological value and the reference morphological value to obtain the first brain morphological index,
a second target morphological value related to the second brain element is obtained based on voxel data corresponding to the second region, and a second brain element and and calculate a related second corrected morphological value, and calculate the second brain morphological indicator based on the second corrected morphological value and the reference morphological value to obtain the second brain morphological indicator. ,
Brain image analysis device.
14. The method of claim 13,
The first morphological value and the second morphological value are values obtained under the first scan condition,
The target brain image is obtained under the first scan condition,
the first calibration parameter is a first parameter for calculating a third morphological value associated with the first brain element obtained under a second scan condition different from the first scan condition based on the first morphological value including,
the second calibration parameter comprises a second parameter for calculating a fourth morphological value associated with the second brain element obtained under the second scan condition based on the second morphological value;
The controller is
calculate the first corrected morphological value, which is a morphological estimate of the first brain element under the second scan condition, based on the first target morphological value and the first correction parameter, wherein the second target morphology configured to calculate the second corrected morphological value that is a morphological estimate of the second brain element under the second scan condition based on the value and the second correction parameter;
Brain image analysis device.
15. The method of claim 14,
the first calibration parameter comprises a first parameter set associated with a first linear function for calculating the third morphological value based on the first morphological value;
the second calibration parameter comprises a second parameter set associated with a second linear function for calculating the fourth morphological value based on the second morphological value;
wherein calculating the first corrected morphological value comprises obtaining the first corrected morphological value based on the first target morphological value and a first parameter set associated with the first linear function,
The controller is
calculate the first corrected morphological value based on the first target morphological value and a first parameter set associated with the first linear function, wherein the second target morphological value and the second linear function are associated with the second target morphological value. configured to calculate the second corrected morphological value based on two parameter sets;
Brain image analysis device.
13. The method of claim 12,
The first region corresponding to the first brain element is located adjacent to the skull region compared to the second region corresponding to the second brain element,
Brain image analysis device.
13. The method of claim 12,
wherein the first brain component is associated with a component that performs a first brain function, and the second brain component is associated with a component that performs a second brain function that is different from the first brain function.
Brain image analysis device.
15. The method of claim 14,
The first scan condition and the second scan condition are a magnetic field strength related to the resolution of a brain image of the brain image acquisition device, a manufacturer of the brain image acquisition device, and a setting parameter related to a shape of a magnetic field generated by the brain imaging device associated with at least one of the setting parameters;
Brain image analysis device.
13. The method of claim 12,
The controller is
The target brain image is obtained from a first object having a first characteristic,
The first correction parameter and the second correction parameter are obtained from a first brain image and a second brain image obtained from a second object having the first characteristic,
wherein the first characteristic is related to the age or sex of the subject;
Brain image analysis device.
13. The method of claim 12,
The segmentation is performed using a neural network provided to obtain a plurality of brain regions corresponding to a plurality of brain elements including the first brain element, the second brain element, and the skull based on the target brain image. felled,
Brain image analysis device.
16. The method of claim 15,
The controller is
The target brain image is preprocessed, and the segmentation is performed based on the preprocessed target brain image, and the first region is preprocessed by a first method, and the second region is different from the first method configured to perform different pretreatments,
Brain image analysis device.
16. The method of claim 15,
The first corrected morphological value is a volume value associated with the first brain element obtained with the target brain image, and the second target morphological value is a volume value associated with the second brain element obtained with the target brain image. is the value,
The first brain morphological index is calculated as the first corrected morphological value with respect to the reference morphological value,
wherein the second brain morphological index is calculated as the second corrected morphological value with respect to the reference morphological value.
Brain image analysis device.

Images