KR102668408B1 - Apparatus and method for automated analysis of atrial wall on cardiac Magnetic Resonance Imaging - Google Patents

Abstract

An atrial wall segmentation device and method for automatically segmenting the atrial wall in cardiac magnetic resonance imaging (MRI) based on an artificial intelligence model are disclosed. The atrial wall segmentation method in cardiac magnetic resonance imaging according to an embodiment of the present invention is: selecting an atrial region from a cardiac magnetic resonance image acquired for the heart of a subject using an artificial intelligence model and extracting an atrial boundary from the atrial region. steps; and determining the atrial wall by expanding from the atrial border to surrounding pixels by using an atrial wall dividing unit.

Description

Atrial wall segmentation device and method in cardiac magnetic resonance imaging {Apparatus and method for automated analysis of atrial wall on cardiac Magnetic Resonance Imaging}

The present invention relates to an atrial wall segmentation device and method in cardiac magnetic resonance imaging (MRI).

The number of arrhythmia patients is increasing due to various social factors such as population aging, changes in eating habits, changes in working environment, and improved convenience of living environment, and the demand for invasive procedures along with drugs in treating arrhythmia is also increasing. In general, for successful arrhythmia treatment through invasive procedures, specific information about the structure and tissue of the heart is required, and cardiac magnetic resonance imaging (MRI) is used for this purpose. In other words, if damage to the heart structure, especially the atrial wall tissue, is identified through cardiac magnetic resonance imaging before the procedure, it will be possible to perform a customized procedure for the patient, thereby increasing the likelihood of success and safety of arrhythmia treatment. However, precise analysis of the thin atrial wall in cardiac magnetic resonance imaging is a task that requires both the analyst's skill and high concentration. As the number of arrhythmia procedures increases, the development of an automatic magnetic resonance image analysis system is also required.

The present invention is intended to provide an atrial wall segmentation device and method for automatically segmenting the atrial wall in cardiac magnetic resonance imaging (MRI) based on an artificial intelligence model.

In addition, the present invention is intended to provide an atrial wall segmentation device and method that can dramatically shorten the atrial wall segmentation and analysis time for cardiac magnetic resonance imaging of a subject.

The atrial wall segmentation method in cardiac magnetic resonance imaging according to an embodiment of the present invention is: selecting an atrial region from a cardiac magnetic resonance image acquired for the heart of a subject using an artificial intelligence model and extracting an atrial boundary from the atrial region. steps; and determining the atrial wall by expanding from the atrial border to surrounding pixels by using an atrial wall dividing unit.

Determining the atrial wall includes: generating an image brightness histogram of pixels surrounding the atrial border; and selecting the atrial wall by analyzing pixel values surrounding the highest image brightness in the image brightness histogram.

The step of generating the image brightness histogram includes: a change in image brightness from a first boundary point corresponding to the atrial region inside the inner border of the atrial border to a second border point extending outside the atrial border toward the outer border of the atrial border. It may include generating an image brightness histogram representing .

The step of determining the atrial wall may include: selecting an image pixel located within a set distance range from the atrial border as the atrial wall.

The artificial intelligence model may be a convolutional neural network model generated by machine learning based on a convolutional neural network.

The method of segmenting the atrial wall in a cardiac magnetic resonance image according to an embodiment of the present invention includes: forming a three-dimensional image by integrating the atrial walls segmented from a plurality of cardiac magnetic resonance images obtained for the heart of the subject; More may be included.

According to an embodiment of the present invention, a computer-readable recording medium on which a program for executing the atrial wall segmentation method in cardiac magnetic resonance imaging is recorded is provided.

The atrial wall segmentation device in cardiac magnetic resonance imaging according to an embodiment of the present invention is an artificial intelligence model configured to select an atrial region from a cardiac magnetic resonance image obtained for the heart of a subject and extract an atrial boundary from the atrial region. ; and an atrial wall dividing unit configured to determine the atrial wall by extending from the atrial border to surrounding pixels.

In one embodiment of the present invention, the atrial wall dividing unit includes: an image brightness histogram generator configured to generate an image brightness histogram of pixels surrounding the atrial border; and an atrial wall selection unit configured to select the atrial wall by analyzing pixel values surrounding the highest image brightness in the image brightness histogram.

The image brightness histogram generator: an image representing a change in image brightness from a first boundary point corresponding to the atrial region inside the inner border of the atrial border to a second border point extending outside the atrial border toward the outer border of the atrial border. It may be configured to generate a brightness histogram.

In another embodiment of the present invention, the atrial wall dividing unit may be configured to select an image pixel located within a set distance range from the atrial border as the atrial wall.

The atrial wall segmentation device in cardiac magnetic resonance imaging according to an embodiment of the present invention: a three-dimensional image that integrates the atrial walls segmented from a plurality of cardiac magnetic resonance images obtained for the heart of the subject to form a three-dimensional image. It may further include a component part.

According to an embodiment of the present invention, an atrial wall segmentation device and method for automatically segmenting the atrial wall in cardiac magnetic resonance imaging (MRI) based on an artificial intelligence model are provided.

In addition, according to an embodiment of the present invention, an atrial wall segmentation device and method that can dramatically shorten the atrial wall segmentation and analysis time for cardiac magnetic resonance images of a subject are provided.

1 is a configuration diagram of an atrial wall segmentation device in cardiac magnetic resonance imaging according to an embodiment of the present invention.
Figure 2 is a configuration diagram of an atrial wall dividing unit constituting an atrial wall dividing device in cardiac magnetic resonance imaging according to an embodiment of the present invention.
Figure 3 is a flowchart of an atrial wall segmentation method in cardiac magnetic resonance imaging according to an embodiment of the present invention.
Figure 4 is a flowchart showing step S40 of Figure 3 in more detail.
Figure 5 is a flowchart showing step S40 of Figure 3 according to another embodiment of the present invention.
Figure 6 is an example diagram showing the results of selecting an atrial region in a cardiac magnetic resonance image using an artificial intelligence model according to an embodiment of the present invention.
Figure 7 is an example diagram showing the results of extracting the atrial boundary from a cardiac magnetic resonance image using an artificial intelligence model according to an embodiment of the present invention.
Figures 8 and 9 are exemplary diagrams for explaining the process of determining the atrial wall according to an embodiment of the present invention.
Figures 10 and 11 are exemplary diagrams to explain in more detail the process of determining the atrial wall according to an embodiment of the present invention.
Figure 12 is an example diagram showing a three-dimensional image of the atrial wall from a cardiac magnetic resonance image of a subject according to an embodiment of the present invention.

The purpose and effects of the present invention, and technical configurations for achieving them, will become clear by referring to the embodiments described in detail below along with the accompanying drawings. In describing the present invention, if it is determined that a detailed description of a known function or configuration may unnecessarily obscure the gist of the present invention, the detailed description will be omitted. The terms described below are defined in consideration of the functions in the present invention, and may vary depending on the intention or custom of the user or operator.

However, the present invention is not limited to the embodiments disclosed below and may be implemented in various different forms. These embodiments are merely provided to ensure that the disclosure of the present invention is complete and to fully inform those skilled in the art of the present invention of the scope of the invention, and that the present invention is defined by the scope of the claims. It just becomes. Therefore, the definition should be made based on the contents throughout this specification.

As used herein, '~unit' refers to a unit that processes at least one function or operation, and may mean, for example, software, FPGA, or hardware components. The functions provided in '~ part' may be performed separately by multiple components, or may be integrated with other additional components. The '~ part' in this specification is not necessarily limited to software or hardware, and may be configured to be in an addressable storage medium, or may be configured to reproduce one or more processors. Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

The atrial wall segmentation device and method in cardiac magnetic resonance imaging according to an embodiment of the present invention selects an atrial region from a cardiac magnetic resonance image obtained for the heart of a subject using an artificial intelligence model, and defines the atrial boundary in the atrial region. The atrial wall is determined by expanding from the extracted atrial border to surrounding pixels.

According to the atrial wall segmentation device and method in cardiac magnetic resonance imaging according to an embodiment of the present invention, the atrial wall can be automatically segmented in cardiac magnetic resonance imaging (MRI) based on an artificial intelligence model, and the subject The atrial wall segmentation and analysis time for cardiac magnetic resonance imaging can be dramatically shortened.

1 is a configuration diagram of an atrial wall segmentation device in cardiac magnetic resonance imaging according to an embodiment of the present invention. Referring to Figure 1, the atrial wall segmentation device 10 in cardiac magnetic resonance imaging according to an embodiment of the present invention includes an input unit 11, a storage unit 12, a display unit 13, an artificial intelligence model 14, It may include an atrial wall division unit 17, a 3D image configuration unit 18, and a control unit 19.

The input unit 11 allows medical experts such as doctors to execute a program to automatically segment the atrial wall (left atrial wall and/or right atrial wall) in cardiac magnetic resonance images based on an artificial intelligence model using the user interface unit. It may be provided to input various commands to do this. The user interface unit may include, for example, an input interface and/or a communication interface such as a keyboard, mouse, touchpad, electronic pen, and voice recognition unit.

The input unit 11 may be provided to receive various data required for automatic atrial wall segmentation of cardiac magnetic resonance images, for example, cardiac magnetic resonance imaging (MRI) of a patient designated by a medical professional using the user interface unit. The input unit 11 may receive data from a cardiac MRI device or read a cardiac magnetic resonance image from the storage unit 12.

The storage unit 12 is a program for learning an artificial intelligence model for segmenting the patient's atrial wall from a cardiac magnetic resonance image, and automatically divides the atrial wall from a cardiac magnetic resonance image based on the learned artificial intelligence model to determine the size of the atrial wall. Programs that perform processes that constitute a 3D image can be stored.

The display unit 13 displays information such as the patient's cardiac MRI image, a two-dimensional image of the atrial wall segmented and extracted from the cardiac MRI image, and a three-dimensional atrial wall image reconstructed from the atrial wall segmented from multiple cardiac MRI images, to the medical terminal. It may be provided to be displayed through a display screen.

The artificial intelligence model 14 can be learned and built using learning data. The artificial intelligence model 14 may select an atrial region from a cardiac magnetic resonance image obtained for the heart of a subject (eg, a patient undergoing cardiac surgery) and extract an atrial boundary from the selected atrial region.

In an embodiment, the artificial intelligence model 14 may be a convolutional neural network model generated by machine learning based on a convolutional neural network (CNN), such as a U-NET model.

In an embodiment, the artificial intelligence model 14 implemented as a U-NET model includes a contracting path that reduces the image size, an expansive path that increases the image size, and a skip connection unit.

The contracting pass sequentially repeats the first convolution process and pooling process (e.g., max pooling process, average pooling process, etc.) through a plurality of contracting layers. It can be done.

The expansive pass can sequentially and repeatedly perform the second convolution process and up sampling process through a plurality of expanding layers.

The skip connection unit can transmit the feature map (convolutional image) extracted from the contracting layer of the contracting pass to the expanding layer of the expansion pass corresponding to the contracting layer.

The artificial intelligence model 14 may include an atrial region selection unit 15 and an atrial boundary extraction unit 16. The atrial region selection unit 15 may select an atrial region from a cardiac magnetic resonance image obtained for the subject's heart. The atrial border extraction unit 16 may extract the atrial border from the atrial region selected by the atrial region selection unit 15.

The atrial wall dividing unit 17 may determine the atrial wall of the object by extending the atrial boundary extracted by the atrial border extraction unit 16 of the artificial intelligence model 14 to surrounding pixels. Figure 2 is a configuration diagram of an atrial wall dividing unit constituting an atrial wall dividing device in cardiac magnetic resonance imaging according to an embodiment of the present invention.

Referring to Figures 1 and 2, the atrial wall dividing unit 17 may include an image brightness histogram generating unit 172 and an atrial wall selecting unit 174. The image brightness histogram generator 172 may generate an image brightness histogram of pixels surrounding the border of the atrium.

The atrial wall selection unit 174 may select the atrial wall by analyzing pixel values surrounding the highest image brightness in the image brightness histogram generated by the image brightness histogram generator 172.

For example, the atrial wall dividing unit 17 detects a change in image brightness from a boundary point corresponding to the atrial region inside the inner border of the extracted atrial border to a border point extending outside the atrial border toward the outer border of the atrial border. You can create a histogram representing the image brightness.

The image brightness histogram generator 172 of the atrial wall dividing unit 17 may generate an image brightness histogram of pixels surrounding the atrial border. For example, the image brightness histogram generator 172 may generate an image brightness histogram showing changes in image brightness from a boundary point corresponding to the inner border of the atrial border toward the outer border of the atrial border.

The atrial wall selection unit 174 may select the atrial wall by analyzing pixel values surrounding the highest image brightness in the image brightness histogram generated by the image brightness histogram generator 172. In one embodiment, the atrial wall selection unit 174 may determine an area having a pixel brightness greater than or equal to a pixel value (pixel brightness) in the area including the atrial border as the atrial wall.

As described above, the atrial wall dividing unit 17 repeats the process of determining the area of the atrial wall based on the image brightness based on the image brightness histogram on the atrial border along the circumference of the atrial border, thereby forming a boundary surrounding the atrium ( The atrial wall corresponding to the border can be determined.

According to another embodiment of the present invention, the atrial wall dividing unit 17 may select image pixels located within a set distance range from the atrial border as the atrial wall. In this case, the atrial wall can be determined to have a constant thickness along the atrial border.

The 3D image constructor 18 may construct a 3D image by integrating the atrial walls segmented from a plurality of cardiac magnetic resonance images obtained for the subject's heart. The atrial wall of the object reconstructed as a 3D image may be displayed on the display screen by the display unit 13.

The control unit 19 includes various components such as an input unit 11, a storage unit 12, a display unit 13, an artificial intelligence model 14, an atrial wall division unit 17, and a 3D image component 18. It may include at least one processor that controls and performs a process of automatically segmenting the atrial wall from the patient's cardiac MRI using an artificial intelligence model.

As described above, the atrial wall segmentation device in cardiac magnetic resonance imaging according to the embodiment of the present invention has an average dice coefficient of 0.94 (range 0.89 - 0.97) of the artificial intelligence model, showing excellent atrial wall segmentation performance. .

According to the atrial wall segmentation device for cardiac magnetic resonance imaging according to an embodiment of the present invention, the time for segmenting and analyzing the atrial wall for cardiac magnetic resonance imaging of a subject can be dramatically shortened.

According to the atrial wall segmentation method in cardiac magnetic resonance imaging according to an embodiment of the present invention, the 3D image analysis time for cardiac magnetic resonance imaging can be reduced to 1/5 of the existing method, which is about 25 minutes. The video analysis time required can be reduced to about 5 minutes.

The atrial wall segmentation device in cardiac magnetic resonance imaging according to an embodiment of the present invention helps support the selection of an effective treatment strategy, especially for patients requiring procedures for arrhythmias, such as electrode ablation requiring image assistance. I can give it.

Figure 3 is a flowchart of an atrial wall segmentation method in cardiac magnetic resonance imaging according to an embodiment of the present invention. Referring to FIGS. 1 and 3 , when a cardiac magnetic resonance image acquired for the heart of an object (e.g., a patient undergoing cardiac surgery) is input (S10), the artificial intelligence model 14 displays the object (e.g., a patient undergoing cardiac surgery). The atrium region can be selected from a cardiac magnetic resonance image obtained for the heart of a patient subject to cardiac surgery (S20).

The artificial intelligence model 14 can extract the atrial boundary from the selected atrial region (S30). In an embodiment, the artificial intelligence model 14 may be a convolutional neural network (CNN)-based machine learning model.

The atrial wall dividing unit 17 may determine the atrial wall of the object by expanding the atrial boundary extracted by the atrial border extraction unit 16 of the artificial intelligence model 14 to surrounding pixels (S40).

Figure 4 is a flowchart showing step S40 of Figure 3 in more detail. Referring to FIGS. 1, 2, and 4, the image brightness histogram generator 172 of the atrial wall dividing unit 17 may generate an image brightness histogram of pixels surrounding the atrial border (S42).

The atrial wall selection unit 174 may select the atrial wall by analyzing pixel values surrounding the highest image brightness in the image brightness histogram generated by the image brightness histogram generator 172 (S44).

For example, the atrial wall dividing unit 17 detects a change in image brightness from a boundary point corresponding to the atrial region inside the inner border of the extracted atrial border to a border point extending outside the atrial border toward the outer border of the atrial border. You can create a histogram representing the image brightness.

The image brightness histogram generator 172 of the atrial wall dividing unit 17 may generate an image brightness histogram of pixels surrounding the atrial border. For example, the image brightness histogram generator 172 may generate an image brightness histogram showing changes in image brightness from a boundary point corresponding to the inner border of the atrial border toward the outer border of the atrial border.

The atrial wall selection unit 174 may select the atrial wall by analyzing pixel values surrounding the highest image brightness in the image brightness histogram generated by the image brightness histogram generator 172. In one embodiment, the atrial wall selection unit 174 may determine an area having a pixel brightness greater than or equal to a pixel value (pixel brightness) in the area including the atrial border as the atrial wall.

As described above, the atrial wall dividing unit 17 repeats the process of determining the area of the atrial wall based on the image brightness based on the image brightness histogram on the atrial border along the circumference of the atrial border, thereby forming a boundary surrounding the atrium ( The atrial wall corresponding to the border can be determined.

Figure 5 is a flowchart showing step S40 of Figure 3 according to another embodiment of the present invention. Referring to FIGS. 1 and 5, the atrial wall segmentation unit 17 determines image pixels located within a set distance range from the atrial border (S46) and selects the determined image pixels as the atrial wall (S48). In this case, the atrial wall can be determined to have a constant thickness along the atrial border.

Referring again to FIGS. 1 and 3, the 3D image constructor 18 may construct a 3D image by integrating the atrial walls divided from a plurality of cardiac magnetic resonance images obtained for the heart of the subject (S50) ). The atrial wall of the object reconstructed as a 3D image may be displayed on the display screen by the display unit 13.

As described above, the atrial wall segmentation method in cardiac magnetic resonance imaging according to the embodiment of the present invention has an average dice coefficient of 0.94 (range 0.89 - 0.97) of the artificial intelligence model, showing excellent atrial wall segmentation performance. .

According to the atrial wall segmentation method in cardiac magnetic resonance imaging according to an embodiment of the present invention, the atrial wall segmentation and analysis time for cardiac magnetic resonance imaging of a subject can be dramatically shortened.

According to the atrial wall segmentation method in cardiac magnetic resonance imaging according to the embodiment of the present invention as described above, the 3D image analysis time for cardiac magnetic resonance imaging can be reduced to 1/5 of the existing method, and Video analysis time, which takes about 25 minutes, can be shortened to about 5 minutes.

The atrial wall segmentation method in cardiac magnetic resonance imaging according to an embodiment of the present invention is particularly helpful in supporting the selection of an effective treatment strategy for patients requiring arrhythmia procedures, such as electrode ablation requiring image assistance. I can give it.

Figures 6 to 12 are exemplary diagrams for explaining an atrial wall segmentation method in cardiac magnetic resonance imaging according to an embodiment of the present invention. Hereinafter, the atrial wall segmentation method in cardiac magnetic resonance imaging according to an embodiment of the present invention will be described in detail with reference to FIGS. 3 to 12.

Figure 6 is an example diagram showing the results of selecting an atrial region in a cardiac magnetic resonance image using an artificial intelligence model according to an embodiment of the present invention. Referring to FIGS. 1, 3, and 6, the artificial intelligence model 14 can select the atrial region 30 from the cardiac magnetic resonance image 20 obtained for the object (S20). In Figure 6, the atrial region 30 is shown in green.

Figure 7 is an example diagram showing the results of extracting the atrial boundary from a cardiac magnetic resonance image using an artificial intelligence model according to an embodiment of the present invention. Referring to FIGS. 1, 3, 7, and 8, the artificial intelligence model 14 can extract the atrial boundary 40 from the selected atrial region 30 (S30). The atrial border 40 extracted by the artificial intelligence model 14 is shown as a green line and arrow in FIG. 7. In an embodiment, the artificial intelligence model 14 may be a convolutional neural network (CNN)-based machine learning model.

Figures 8 and 9 are exemplary diagrams for explaining the process of determining the atrial wall according to an embodiment of the present invention. Referring to FIGS. 1, 3, 8, and 9, the atrial wall division unit 17 divides the surrounding pixels from the atrial border 40 extracted by the atrial border extraction unit 16 of the artificial intelligence model 14. The atrial wall 50 of the object can be determined by expanding (shown by a double-headed arrow in FIG. 9) (S40).

The atrial wall dividing unit 17 may generate an image brightness histogram 60 of pixels surrounding the atrial border 40 and determine the atrial wall 50 based on the brightness information of the image brightness histogram 60.

The atrial wall dividing portion 17 extends from the first border point 41, 45 corresponding to the atrial region inside the inner border of the atrial border 40 toward the outer border of the atrial border 40. An image brightness histogram 60 showing changes in image brightness can be generated up to the second boundary points 43 and 47 extending to .

Figures 10 and 11 are exemplary diagrams to explain in more detail the process of determining the atrial wall according to an embodiment of the present invention. Referring to FIGS. 1, 2, 4, 10, and 11, the image brightness histogram generator 172 of the atrial wall dividing unit 17 generates an image brightness histogram 60 of pixels surrounding the atrial border 40. can be created (S42).

The image brightness histogram generator 172 generates an image brightness display that represents the change in image brightness from the first boundary points 42 and 46 corresponding to the inner border of the atrial border 40 toward the outer border of the atrial border 40. A histogram 60 can be created.

The atrial wall selection unit 174 may select the atrial wall 50 by analyzing pixel values surrounding the highest image brightness in the image brightness histogram 60 generated by the image brightness histogram generator 172 (S44).

In one embodiment, the atrial wall selection unit 174 may determine an area having a pixel brightness greater than or equal to the pixel value (pixel brightness) of the first boundary points 42 and 46 as the atrial wall 50. In the example shown in FIG. 11, the pixel value of the first boundary points 42 and 46 is “215”, so the pixel value on the outer edge of the atrial boundary 40 from the first boundary points 42 and 46 is “215”. Up to the second boundary points 44 and 48 may be determined as the area corresponding to the atrial wall 50.

As described above, the atrial wall dividing unit 17 repeats the process of determining the area of the atrial wall 50 based on the image brightness based on the image brightness histogram 60 along the circumference of the atrial border 40, The atrial wall 50 corresponding to the boundary (rim) surrounding the atrium can be determined.

Meanwhile, unlike the embodiment shown in FIGS. 6 to 11, the atrial wall dividing unit 17 may determine image pixels located within a set distance range from the atrial border and select the determined image pixels as the atrial wall. In this case, the atrial wall can be determined to have a constant thickness along the atrial border.

Figure 12 is an example diagram showing a three-dimensional image of the atrial wall from a cardiac magnetic resonance image of a subject according to an embodiment of the present invention. Referring to FIGS. 1, 3, and 12, the 3D image constructor 18 reconstructs the 3D image 70 by integrating the atrial walls divided from a plurality of cardiac magnetic resonance images obtained for the heart of the subject. You can do it (S50). The atrial wall of the object reconstructed in the 3D image 70 may be displayed on the display screen by the display unit 13.

According to the atrial wall segmentation method in cardiac magnetic resonance imaging according to the embodiment of the present invention as described above, the 3D image analysis time for cardiac magnetic resonance imaging can be reduced to 1/5 of the existing method, and Video analysis time, which takes about 25 minutes, can be shortened to about 5 minutes.

In addition, the atrial wall segmentation method in cardiac magnetic resonance imaging according to an embodiment of the present invention has an average dice coefficient of the artificial intelligence model of 0.94 (range 0.89 - 0.97), showing excellent atrial wall segmentation performance.

The atrial wall segmentation method in cardiac magnetic resonance imaging according to an embodiment of the present invention will help support the selection of an effective treatment strategy for patients requiring arrhythmia procedures, for example, electrode ablation requiring image support. You can.

The embodiments described above may be implemented with hardware components, software components, and/or a combination of hardware components and software components. For example, the devices, methods, and components described in the embodiments may include, for example, a processor, a controller, an Arithmetic Logic Unit (ALU), a Digital Signal Processor, a microcomputer, and a Field Programmable Gate (FPGA). It may be implemented using one or more general-purpose computers or special-purpose computers, such as an array, PLU (Programmable Logic Unit), microprocessor, or any other device that can execute and respond to instructions.

The processing device may execute an operating system and one or more software applications that run on the operating system. Additionally, a processing device may access, store, manipulate, process, and generate data in response to the execution of software. For ease of understanding, a single processing device may be described as being used; however, those skilled in the art will understand that a processing device may include multiple processing elements and/or multiple types of processing elements. You will understand that it can be included.

For example, a processing device may include a plurality of processors or one processor and one controller. Additionally, other processing configurations, such as parallel processors, are also possible. Software may include a computer program, code, instructions, or a combination of one or more of these, and may configure a processing unit to operate as desired, or may be processed independently or collectively. You can command the device.

Software and/or data may be used on any type of machine, component, physical device, virtual equipment, computer storage medium or device to be interpreted by or to provide instructions or data to a processing device. , or may be permanently or temporarily embodied in a transmitted signal wave. Software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium. Computer-readable media may include program instructions, data files, data structures, etc., singly or in combination. Program instructions recorded on the medium may be specially designed and configured for the embodiment or may be known and available to those skilled in the art of computer software.

Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CDROMs and DVDs, and ROM, RAM, and flash memory. Includes hardware devices specifically configured to store and execute program instructions, such as: Examples of program instructions include machine language code, such as that produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

The above description is merely an illustrative explanation of the technical idea of the present invention, and various modifications and variations will be possible to those skilled in the art without departing from the essential characteristics of the present invention. Accordingly, the embodiments disclosed in the present invention are not intended to limit the technical idea of the present invention, but are for illustrative purposes, and the scope of the technical idea of the present invention is not limited by these embodiments. The scope of protection of the present invention should be interpreted in accordance with the claims below, and all technical ideas within the equivalent scope should be construed as being included in the scope of rights of the present invention.

10: Atrial wall segmentation device in cardiac magnetic resonance imaging
11: input unit
12: storage unit
13: display part
14: Artificial intelligence model
15: Atrial region selection unit
16: Atrial border extraction unit
17: Atrial wall division
18: 3D image component
19: Control unit
20: Cardiac magnetic resonance imaging
30: Atrium area
40: Atrium border
42: first boundary point
44: Second boundary point
50: Atrial wall
60: Image brightness histogram
70: 3D video
172: Image brightness histogram generation unit
174: Atrial wall selection section

Claims (13)

Selecting an atrial region from a cardiac magnetic resonance image obtained for the heart of an object using an artificial intelligence model and extracting an atrial boundary from the atrial region; and
A step of determining the atrial wall by extending from the atrial border to surrounding pixels by an atrial wall dividing unit,
The steps for determining the atrial wall are:
generating an image brightness histogram of pixels surrounding the atrial border; and
A step of selecting an atrial wall by analyzing pixel values surrounding the highest image brightness in the image brightness histogram,
The steps for generating the image brightness histogram are:
Generate an image brightness histogram showing the change in image brightness from a first boundary point corresponding to the atrial region of the inner border of the atrial border to a second border point extending outside the atrial border toward the outer border of the atrial border, wherein the atrium Generating the image brightness histogram along the perimeter of the boundary,
The steps for selecting the atrial wall are:
In the image brightness histogram generated from the first boundary point to the second boundary point, determining an area having a pixel brightness greater than or equal to the pixel value of the first boundary point as the area of the atrial wall, the atrium in a cardiac magnetic resonance image. How to split a wall. delete delete delete According to paragraph 1,
The artificial intelligence model is a convolutional neural network model generated by machine learning based on a convolutional neural network. An atrial wall segmentation method in cardiac magnetic resonance imaging. According to paragraph 1,
Atrial wall in a cardiac magnetic resonance image further comprising: integrating the segmented atrial walls from the plurality of cardiac magnetic resonance images obtained for the heart of the object to form a three-dimensional image by a three-dimensional image constructor; How to split. A computer-readable recording medium on which a program for executing the atrial wall segmentation method in cardiac magnetic resonance imaging according to any one of claims 1, 5, and 6 is recorded. An artificial intelligence model configured to select an atrial region from a cardiac magnetic resonance image obtained for the heart of an object and extract an atrial boundary from the atrial region; and
An atrial wall dividing unit configured to determine the atrial wall by extending from the atrial border to surrounding pixels,
The atrial wall division:
an image brightness histogram generator configured to generate an image brightness histogram of pixels surrounding the atrium border; and
An atrial wall selection unit configured to select an atrial wall by analyzing pixel values surrounding the highest image brightness in the image brightness histogram,
The image brightness histogram generator:
Generating an image brightness histogram showing the change in image brightness from a first boundary point corresponding to the atrial region of the inner border of the atrial border to a second border point extending outside the atrial border toward the outer border of the atrial border, wherein the atrium Configured to generate the image brightness histogram along the perimeter of the boundary,
The atrial wall selection unit is configured to determine, as the area of the atrial wall, an area having a pixel brightness greater than or equal to the pixel value of the first boundary point in the image brightness histogram generated from the first boundary point to the second boundary point. Atrial wall splitting device in . delete delete delete According to clause 8,
The artificial intelligence model is an atrial wall segmentation device in cardiac magnetic resonance imaging, which is a convolutional neural network model generated by machine learning based on a convolutional neural network. According to clause 8,
An atrial wall segmentation device in a cardiac magnetic resonance image, further comprising a 3D image composition unit that configures a 3D image by integrating the atrial walls segmented from the plurality of cardiac magnetic resonance images obtained for the heart of the subject.

Images