KR102416928B1 - Active bleeding detection system and method using abdominal CT(computed tomography) image analysis - Google Patents

Abstract

The present invention relates to an active bleeding detection system and method using abdominal computed tomography (CT) image analysis. According to the present invention, the active bleeding detection method using abdominal CT image analysis includes: a step of receiving, by an image inputting unit, receiving an abdominal CT image from the outside; a step of arranging, by an image arrangement unit, images before and after injecting a contrast medium in the received CT image; a step of using, by an image segmenting unit, a spine section segmenting model to segment a spine section into a plurality of sections and using a digestive organ segmenting model to segment a digestive organ into a plurality of sections respectively in response to the plurality of segmented sections; a step of inspecting, by a bleeding detecting unit, bleeding of the digestive organ segmented into the plurality of sections and determining whether bleeding is detected according to bleeding inspection; a step of forming, by a bleeding shape forming unit, a bleeding shape if bleeding is detected; a step of determining, by the bleeding detecting unit, whether bleeding is detected from a plurality of sites if bleeding is detected; a step of designating, by an emergency care priority designating unit, priority of emergency care for bleeding if bleeding is detected from the plurality of sites; a step of visualizing, by a bleeding location and shape displaying unit, the location and shape of bleeding detected from the plurality of sites to be displayed on a screen; and a step of determining, by a control unit, whether to end the CT image analysis and end the analysis process if it is determined to end the CT image analysis. The present invention has an effect of helping emergency measures or setting of treatment plans directly and indirectly.

Description

Active bleeding detection system and method using abdominal computed tomography (CT) image analysis

The present invention relates to a system and method for detecting active bleeding using abdominal CT image analysis, and more particularly, to an active bleeding using abdominal CT image analysis that can accurately detect the bleeding location of a bleeding patient inside the digestive tract through abdominal CT image analysis It relates to a bleeding detection system and method.

There are cases where there are bleeding findings on CT in the emergency room of the hospital, but other doctors who are not familiar with CT images may not recognize it accurately in a timely manner. In addition, the diagnosis may be delayed because CT images cannot be checked in time due to the rush of emergency patients. As a result, in these situations, accurate diagnosis and first aid are not performed quickly for emergency patients, which may result in a more serious condition for the patient or, in the worst case, death.

Therefore, there is an urgent need for a system that accurately detects the bleeding location of a bleeding patient in the digestive tract through abdominal CT image analysis and displays it on the screen, so that medical staff can quickly identify it and provide emergency treatment.

Korean Patent Publication No. 10-2017-0086311 (2017.07.26.)

The present invention was created in consideration of the above, and provides a system and method for detecting active bleeding using abdominal CT image analysis that can accurately detect the bleeding location of a bleeding patient inside the digestive tract through abdominal CT image analysis. has its purpose in

In order to achieve the above object, an active bleeding detection system using abdominal CT image analysis according to the present invention,

an image input unit for receiving an abdominal CT image from the outside;

an image alignment unit for aligning images before and after injection of a contrast medium in the CT image received through the image input unit;

In the aligned image, the vertebral part is divided into a plurality of sections using the spine section segmentation model, and the digestive system is divided into a plurality of sections by using the digestive organ segmentation model in correspondence with the plurality of sections of the divided spine. with installments;

a bleeding detection unit for examining bleeding in the digestive tract divided into the plurality of sections and determining whether bleeding is detected according to the bleeding test;

a bleeding shape configuration unit that configures a shape of the detected bleeding when the bleeding is detected in the determination of the bleeding detection unit;

an emergency treatment priority designation unit for designating a bleeding emergency treatment priority when bleeding is detected at a plurality of locations in the determination of the bleeding detection unit;

a bleeding location/shape display unit that visualizes the location and shape of the bleeding detected in the plurality of locations and displays it on a screen; and

By controlling the status check and operation of the image input unit, image alignment unit, image segmentation unit, bleeding detection unit, bleeding shape configuration unit, first aid ranking unit, and bleeding location/shape display unit, and determining whether the analysis of the CT image is finished, It is characterized in that it includes a control unit that returns the analysis process to the step of receiving an image by the image input unit if it is not finished, and terminates the analysis process if it is finished.

Here, the CT image aligned by the image alignment unit is a pre-phase image, an arterial-phase image after contrast medium injection, and a second set time after contrast medium injection (Portal- phase) image and a third set time elapsed (Delayed-phase) image after injection of the contrast agent may be included.

In addition, when the image segmentation unit divides the spine into a plurality of sections, the spine may be divided into three sections: thoracic vertebrae, lumbar vertebrae, and sacral/caudal vertebrae.

In addition, when the image segmentation unit divides the digestive tract into a plurality of sections using the digestive organ segmentation model, the digestive organ segmentation model corresponds to the segmented vertebral segment in response to the thoracic segment segmentation model, the lumbar segment segmentation model, the sacrum / It may include a caudal segmentation model.

At this time, when the image segmentation unit divides the digestive tract into a plurality of sections using the digestive organ segmentation model, the digestive system is divided into the esophagus, stomach, ascending/descending colon, and transverse colon segment using the thoracic segment segmentation model. , can be divided into transverse colon and ascending/descending colon section using the lumbar segment segmentation model, and rectal and s-colon segment using the sacral/caudal segment segmentation model, respectively.

In addition, when the bleeding detection unit determines whether bleeding is detected according to the bleeding test of the digestive tract, it is possible to determine whether bleeding is detected by comparing the images before and after injection of the contrast medium using the digestive tract bleeding detection model.

In this case, the digestive system bleeding detection model may include a thoracic region bleeding detection model, a lumbar region bleeding detection model, and a sacral/caudal region bleeding detection model.

In addition, when the bleeding shape component configures the bleeding shape, it is possible to configure the bleeding shape in 3D.

In addition, when the first aid priority designation unit assigns the first aid treatment priority to bleeding when bleeding is detected in multiple locations, when multiple bleeding is found in the images 20 seconds, 60 seconds, and 100 seconds or more after contrast medium injection , it is possible to prioritize first aid treatment by calculating the amount of bleeding or the bleeding rate at the site where bleeding is found.

In addition, preferably, it may further include a communication unit for transmitting/receiving data or information with an external device.

In addition, in order to achieve the above object, the active bleeding detection method using abdominal CT image analysis according to the present invention,

Activity based on an active bleeding detection system using abdominal CT image analysis including an image input unit, an image alignment unit, an image segmentation unit, a bleeding detection unit, a bleeding shape configuration unit, a first aid ranking unit, a bleeding location/shape display unit, and a control unit A method for detecting bleeding comprising:

a) the image input unit receiving an abdominal CT image from the outside;

b) aligning, by the image aligning unit, images before and after injection of a contrast medium in the CT image received through the image input unit;

c) The image segmentation unit divides the vertebral part into a plurality of sections using the vertebral section segmentation model in the aligned image, and divides the digestive system into a plurality of sections using the digestive system segmentation model in correspondence with a plurality of sections of the divided spine dividing each into

d) examining, by the bleeding detection unit, bleeding in the digestive system divided into the plurality of sections;

e) determining, by the bleeding detection unit, whether bleeding is detected according to the bleeding test of the digestive organs;

f) if bleeding is detected in the determination of step e), configuring the bleeding shape by the bleeding shape component;

g) if bleeding is detected in the determination of step e), determining by the bleeding detection unit whether bleeding is detected at a plurality of locations;

h) if bleeding is detected in a plurality of locations in the determination of step g), designating the bleeding emergency treatment priority by the emergency treatment priority designation unit;

i) the bleeding location/shape display unit visualizes the location and shape of the bleeding detected in the plurality of locations and displays it on a screen; and

j) the control unit determines whether the analysis of the CT image is finished, if not, returns the analysis process to step a), and if it is finished, the analysis process is terminated.

Here, in step a), the abdominal CT image is a pre-phase image, an arterial-phase image after contrast medium injection, and a second set time after contrast medium injection (Portal- phase) image and a third set time elapsed (Delayed-phase) image after injection of the contrast agent may be included.

In addition, when the image segmentation unit divides the vertebrae into a plurality of sections in step c), the vertebrae may be divided into three sections: thoracic, lumbar, and sacral/caudal vertebrae.

In addition, the digestive system segmentation model in step c) may include a thoracic segment segmentation model, a lumbar segment segmentation model, and a sacral/caudal segment segmentation model corresponding to the divided vertebral segments.

At this time, when the image segmentation unit divides the digestive tract into a plurality of sections using the digestive organ segmentation model, the digestive system is divided into the esophagus, stomach, ascending/descending colon, and transverse colon segment using the thoracic segment segmentation model. , can be divided into transverse colon and ascending/descending colon section using the lumbar segment segmentation model, and rectal and s-colon segment using the sacral/caudal segment segmentation model, respectively.

In addition, in step e), when the bleeding detection unit determines whether bleeding is detected according to the bleeding test of the digestive tract, it is determined whether bleeding is detected by comparing the images before and after the contrast medium injection using the digestive tract bleeding detection model. can be discerned.

In this case, the digestive system bleeding detection model may include a thoracic region bleeding detection model, a lumbar region bleeding detection model, and a sacral/caudal region bleeding detection model.

Also, in step f), when the bleeding shape component configures the bleeding shape, the shape of the bleeding may be configured in 3D.

In addition, when the first aid priority designation unit in step h) detects bleeding at multiple locations, in assigning the bleeding first aid priority priority, a plurality of If bleeding is found, the first aid can be prioritized by calculating the amount of bleeding or the bleeding rate at the site where the bleeding was found.

According to the present invention as described above, by detecting the exact location of bleeding in the CT image of a bleeding patient inside the digestive tract, there is an effect that can directly/indirectly help a doctor to take emergency measures or set a treatment direction.

In addition, there is an advantage that the diagnosis can be made without delay for bleeding patients at all hospitals and clinics where radiologists do not reside.

In addition, it has the advantage of being able to represent the shape of the bleeding in 3D and to calculate the amount of bleeding and the bleeding rate.

In addition, when bleeding occurs in multiple locations, there is an advantage in that the priority required for emergency treatment can be designated and indicated by calculating the amount of bleeding or the bleeding rate.

1 is a diagram schematically showing the configuration of an active bleeding detection system using abdominal CT image analysis according to the present invention.
2 is a flowchart illustrating an execution process of a method for detecting active bleeding using abdominal CT image analysis according to the present invention.
3 is a diagram illustrating a 3D abdominal CT image as a 2D slice coronal, sagittal, and axial plane images.
4 is a view showing images before and after injection of a contrast medium in the case of descending colonic bleeding of the large intestine.
5 is a view showing images before and after injection of a contrast medium in the case of ascending colonic bleeding of the large intestine.
6 is a view showing an axial plane image of segmentation of the spine and major digestive organs.
7 is a diagram illustrating a sagittal plane image divided into three sections based on a spinal region before contrast medium injection.
8 is a diagram illustrating a state in which regions are segmented using a digestive system segmentation model for each vertebral section of an axial plane image before and after contrast medium injection.
9 is a view showing the state of bleeding detection for each spinal section of an axial plane image before and after injection of a contrast medium.
10 is a view showing a state in which the digestive organs and bleeding regions are marked on an axial plane image before and after injection of a contrast medium.
11 is a diagram illustrating a state in which the digestive organs and bleeding regions are marked according to bitwise AND and bitwise OR in axial plane images before and after injection of a contrast medium.
12 is a diagram illustrating a state in which positions are marked on axial, coronal, and sagittal images according to a bleeding detection result of an image after injection of a contrast medium.
13 is a diagram illustrating the configuration of a bleeding shape in 3D after detecting bleeding in an image after contrast medium injection.
14 is a diagram illustrating the calculation of the amount of bleeding based on an image after injection of a contrast medium.
15 is a diagram illustrating an outline for designating emergency treatment priorities by calculating the amount of bleeding when multiple bleeding is found in an image after injection of a contrast medium.
16 is a diagram illustrating an interval and a thickness of 1 pixel of an axial image in a slice.

The terms or words used in the present specification and claims should not be construed as limited to their ordinary or dictionary meanings, and the inventor may appropriately define the concept of the term in order to best describe his invention. Based on the principle, it should be interpreted as meaning and concept consistent with the technical idea of the present invention.

Throughout the specification, when a part "includes" a certain element, it means that other elements may be further included, rather than excluding other elements, unless otherwise stated. In addition, terms such as “…unit”, “…group”, “module”, and “device” described in the specification mean a unit that processes at least one function or operation, which is hardware or software or a combination of hardware and software. can be implemented as

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Here, before describing the embodiments of the present invention, the terms used in the present invention will be first described.

In the present invention, "active bleeding" mostly means "arterial bleeding".

1 is a diagram schematically showing the configuration of an active bleeding detection system using abdominal CT image analysis according to an embodiment of the present invention.

Referring to FIG. 1 , an active bleeding detection system 100 using abdominal CT image analysis according to the present invention includes an image input unit 110 , an image alignment unit 120 , an image division unit 130 , a bleeding detection unit 140 , It is configured to include a bleeding shape configuration unit 150 , a first aid ranking unit 160 , a bleeding location/shape display unit 170 , and a control unit 180 .

The image input unit 110 receives an abdominal CT image from the outside (eg, an integrated server that stores and manages CT image data of a hospital or clinic, a personal terminal of a doctor, or a CT imaging device). Here, the image input unit 110 may receive a 3D abdominal CT image, and the bleeding location/shape display unit 170, which will be described later, displays the received 3D abdominal CT image as shown in FIG. 3 , a 2D slice Coronal. A plane image (a), a sagittal plane image (b), and an axial plane image (c) can be converted and displayed.

The image alignment unit 120 aligns the images before and after injection of the contrast medium in the CT image received through the image input unit 110 . At this time, the CT image aligned by the image alignment unit 120 is, as shown in FIGS. 4 and 5 , a pre-phase image (a) after contrast medium injection and a first set time after contrast medium injection (eg For example, after 20 seconds) (Arterial-phase) image (b) and after a second set time (eg, after 60 seconds) after contrast medium injection (Portal-phase) image (c) and the third after contrast medium injection It may include an image after a set time has elapsed (Delayed-phase). Here, after 20 seconds and after 60 seconds are examples, and are not necessarily limited to such time values, and may be set to other time values according to an applied object or situation. 4 shows images before and after contrast medium injection in the case of descending colonic bleeding, and FIG. 5 shows images before and after contrast medium injection in the case of ascending colonic bleeding.

The image segmentation unit 130 divides the spinal segment into a plurality of sections using the spinal segment segmentation model 130a in the image aligned by the image arrangement unit 120, and digests the segments corresponding to the segments of the segmented spine. Each of the digestive organs is divided into a plurality of sections using the organ division model 130b. Here, when the image dividing unit 130 divides the spine into a plurality of sections, as shown in FIG. 6 , the spinal section may be divided into three sections: thoracic, lumbar, and sacral/coccyx. In addition, when the image segmentation unit 130 divides the digestive system into a plurality of sections using the digestive organ segmentation model 130b, the digestive organ segmentation model 130b is It may include a thoracic segment segmentation model, a lumbar segment segmentation model, and a sacral/caudal segment segmentation model corresponding to the segmented spine segment. At this time, when the image segmentation unit 130 divides the digestive system into a plurality of sections using the digestive organ segmentation model 130b, respectively, the digestive system uses the thoracic segment segmentation model to divide the digestive system into the esophagus, stomach, and ascending · It can be divided into descending colon and transverse colon sections, transverse colon and ascending/descending colon sections using the lumbar section segmentation model, and rectum and s-colon sections using the sacral/caudal segment segmentation model.

The bleeding detection unit 140 examines the bleeding of the digestive tract divided into the plurality of sections, and determines whether bleeding is detected according to the bleeding test. Here, when the bleeding detection unit 140 determines whether bleeding is detected according to the bleeding test of the digestive tract, as shown in FIG. 9 , images before and after injection of the contrast medium are compared using the digestive tract bleeding detection model. By doing so, it is possible to determine whether or not bleeding is detected. In this case, the digestive system bleeding detection model may include a thoracic region bleeding detection model, a lumbar region bleeding detection model, and a sacral/caudal region bleeding detection model.

Here, we will try to add a little more explanation in relation to the Arterial-phase, Portal-phase, and Delayed-phase images.

If bleeding is detected from the arterial-phase (20-40 sec) image, it can be considered as possible "arterial" bleeding (this is the case in most cases if there is a contrast medium leak on CT), and the portal-phase (60-90 sec) image If bleeding is detected from a “portal” bleeding, it can be considered as a possibility of “venous” bleeding if bleeding is detected from a Delayed-phase (more than 100 seconds) image.

When bleeding is detected by the bleeding detection unit 140 , the bleeding shape configuration unit 150 configures the detected bleeding shape as shown in FIG. 13 . Here, when the bleeding shape component 150 configures the bleeding shape, the bleeding shape may be configured in 3D. Here, a three-dimensional connected component of the detected blood outflow is also analyzed. That is, each bleeding is classified by labeling the bleeding by applying the 3D connection component analysis to the bleeding detected in the axial slice.

The first aid priority designation unit 160 assigns the bleeding emergency treatment priority order if bleeding is detected in a plurality of locations in the determination of the bleeding detection unit 140 . Here, as shown in FIG. 15 , the first aid ranking unit 160 determines where bleeding is found when a plurality of bleeding is found in the images 20 seconds, 60 seconds, and 100 seconds or more after contrast medium injection. The first aid priority (①,②) can be specified by calculating the bleeding amount or bleeding rate. Here, the calculation of the bleeding amount or bleeding rate from the image after contrast medium injection will be described in a little more detail.

In the DICOM (Digital Imaging and Communications in Medicine) tag of the CT image, as shown in FIG. 16 , the horizontal and vertical pixel spacing of 1 pixel of the axial image and the slice thickness are stored in mm units. . At this time, if the volume of 1 pixel is converted into units, 3 cubic millimeters (mm ³ ) = 0.003 milliliters (ml). Based on this, B _Ai (the amount of bleeding (ml) for the i-th blood spillage detected in the arterial-phase), B _Pi (the amount of bleeding (ml) for the ith blood outflow detected in the portal-phase), B _Di (the amount of bleeding (ml) for the i-th blood outflow detected in the delayed-phase) can be expressed by the following equations.

B _Ai = k × V _Ai

B _Pi = k × V _Pi

B _Di = k × V _Di

where k is the ml value corresponding to the mm ³ volume of 1 pixel of the image, V _Ai is the number of pixels for the i-th blood outflow detected in the arterial-phase, and V _Pi is the i-th blood outflow detected in the portal-phase. The number of pixels, V _Di represents the number of pixels for the i-th blood outflow detected in the Delayed-phase, respectively.

In addition, the bleeding rate can be calculated by calculating the amount of bleeding per second (ml/sec) between phases for the same bleeding. This can be expressed as a formula relation as follows.

The bleeding location/shape display unit 170 visualizes the location and shape of the bleeding detected in the plurality of locations and displays it on the screen. The bleeding position/shape display unit 170 may be configured as a general LCD monitor.

The control unit 180 includes the image input unit 110 , the image alignment unit 120 , the image division unit 130 , the bleeding detection unit 140 , the bleeding shape configuration unit 150 , the first aid ranking unit 160 , the bleeding Controls the state check and operation of the position/shape display unit 170, determines whether the analysis of the CT image is finished, and if not, returns the analysis process to the step of receiving the image by the image input unit 110, and ends Terminates the analysis process. Such a control unit 180 may be composed of a microprocessor, a microcontroller, or the like.

Here, the active bleeding detection system 100 using abdominal CT image analysis according to the present invention having the above configuration preferably further includes a communication unit 190 for transmitting/receiving data or information with an external device or network. may include

Here, the active bleeding detection system 100 using abdominal CT image analysis according to the present invention as described above may be configured as a single computer system (eg, a general desktop computer) in some cases.

Hereinafter, a method for detecting active bleeding using abdominal CT image analysis based on the active bleeding detection system using abdominal CT image analysis according to the present invention having the above configuration will be described.

2 is a flowchart illustrating an execution process of a method for detecting active bleeding using abdominal CT image analysis according to an embodiment of the present invention.

Referring to FIG. 2 , the active bleeding detection method using abdominal CT image analysis according to the present invention includes the image input unit 110 , the image alignment unit 120 , the image division unit 130 , and the bleeding detection unit 140 as described above. ), bleeding shape component 150, first aid ranking unit 160, bleeding location / shape display unit 170, and an active bleeding detection system 100 using abdominal CT image analysis including a control unit 180 based on As a method for detecting active bleeding, first, the image input unit 110 receives an abdominal CT image from the outside (step S201). Here, the abdominal CT image is a pre-phase image, an Arterial-phase image after a first set time (for example, 20 seconds) after contrast agent injection, and a second set time (e.g. For example, it may include an image after 60 seconds) elapses (Portal-phase).

When an abdominal CT image is input through the image input unit 110 , the image alignment unit 120 aligns the images before and after injection of the contrast medium in the CT image received through the image input unit 110 (step S202 ). Here, the CT image aligned by the image alignment unit 120 is, as shown in FIGS. 4 and 5 , a pre-phase image (a) after contrast medium injection and a first set time after contrast medium injection (eg, For example, after 20 seconds (Arterial-phase) image (b), after a second set time has elapsed (for example, after 60 seconds) after contrast medium injection (Portal-phase) image (c), and after contrast medium injection, the third It may include an image after a set time has elapsed (Delayed-phase).

When the image alignment is completed by the image alignment unit 120 in this way, the image division unit 130 divides the spine into a plurality of sections using the spine section division model 130a in the aligned image, and divides The digestive system is divided into a plurality of sections by using the digestive system segmentation model 130b in correspondence with the plurality of sections of the vertebrae (step S203). Here, when the image dividing unit 130 divides the spine into a plurality of sections, as shown in FIG. 6 , the image divider 130 may divide the spine into three sections: thoracic, lumbar, and sacral/coccyx. In addition, as shown in FIG. 8 , the digestive system segmentation model 130b may include a thoracic segment segmentation model, a lumbar segment segmentation model, and a sacral/caudal segment segmentation model corresponding to the segmented vertebrae segments. At this time, when the image segmentation unit 130 divides the digestive system into a plurality of sections using the digestive system division model, the digestive system is divided into the esophagus, stomach, ascending and descending colon using the thoracic segment division model, As a transverse colon section, it can be divided into a transverse colon and ascending/descending colon section using the lumbar segment segmentation model, and a rectum and s-colon segment using the sacral/coccid segment segmentation model, respectively.

In this way, when the digestive system is divided into a plurality of sections by the image dividing unit 130, the bleeding detection unit 140 inspects the bleeding in the digestive organs divided into the plurality of sections (step S204), and according to the bleeding test It is determined whether bleeding has been detected (step S205). Here, when the bleeding detection unit 140 determines whether bleeding is detected according to the bleeding test of the digestive tract, it is determined whether bleeding is detected by comparing the images before and after the contrast medium injection using the digestive tract bleeding detection model. can In this case, the digestive system bleeding detection model, as shown in FIG. 9 , may include a bleeding detection model in the thoracic region, a bleeding detection model in the lumbar region, and a bleeding detection model in the sacral/caudal region.

If bleeding is detected in the determination in step S205, as shown in FIG. 13, the bleeding shape configuration unit 150 configures a bleeding shape (step S206). Here, when the bleeding shape component 150 configures the bleeding shape, the bleeding shape may be configured in 3D. In addition, if no bleeding is detected in the determination of step S205, the analysis process proceeds to step S210.

In addition, when bleeding is detected in the determination in step S205, the bleeding detection unit 140 determines whether bleeding is detected in a plurality of locations (step S207). In this determination, if no bleeding is detected at a plurality of locations, the analysis process advances to step S209.

And, if bleeding is detected in a plurality of locations in the determination of step S207, the first aid treatment priority designating unit 160 assigns the bleeding first aid treatment priority order (step S208). Here, when the first aid priority designation unit 160 determines that bleeding is detected at multiple locations, in designating the bleeding emergency treatment priority, as shown in FIG. 15 , after injection of the contrast medium 20 seconds, 60 seconds, If multiple bleeding is detected in the image for more than 100 seconds, the first aid priority (①,②) can be designated by calculating the amount of bleeding or the bleeding rate at the site where the bleeding was found.

As described above, after the first aid priority is designated by the first aid ranking unit 160, the bleeding location/shape display unit 170 visualizes the location and shape of the bleeding detected in the plurality of locations and displays it on the screen. (step S209).

Thereafter, the controller 180 determines whether the analysis of the CT image is finished, and if not, returns the analysis process to the step S201, and if it is finished, the analysis process is terminated.

Here, we will try to add a little more description in relation to the system and method of the present invention as described above.

7 is a diagram illustrating a sagittal plane image divided into three sections based on a spinal region before contrast medium injection.

Referring to FIG. 7 , when a sagittal plane image is input before contrast medium injection, the image segmentation unit 130 divides the spine into a plurality of (here, three) sections (eg, thoracic segment, It is divided into lumbar vertebrae and sacral/caudal vertebrae), and the axial slice is mapped to the divided vertebral segments. At this time, the spinal section division model 130a, as shown, learns in advance for the correct answer data of the spine section division, and divides the spine part into a plurality of (here, three) sections (for example, It will be divided into thoracic vertebrae, lumbar vertebrae, and sacral/caudal vertebrae).

8 is a diagram illustrating a state in which regions are segmented using a digestive system segmentation model for each vertebral section of an axial plane image before and after contrast medium injection.

8, (a) is an axial image of the lumbar region before contrast agent injection, (b) is an axial image of the lumbar region 20 seconds after contrast agent injection, (c) is an axial image of the lumbar region 60 seconds after contrast agent injection It is an axial image. When the axial image of the lumbar segment of (a), (b), and (c) as above is input, the lumbar segment is divided among the digestive system segmentation models (thoracic segment segmentation model, lumbar segment segmentation model, sacral vertebrae segment segmentation model). The digestive system region of the lumbar vertebrae is output using the model. In the same principle, when the axial image of the thoracic segment is input, the digestive system region of the thoracic segment is output using the thoracic segment segmentation model. The digestive system area of the sacral/coccyx section is output. At this time, similarly, each digestive system segmentation model for each vertebral segment is divided into the thoracic vertebrae, lumbar vertebrae, and sacral/coccyx segments (ie, thoracic vertebra: esophagus, stomach, ascending/descending colon, lumbar vertebra: transverse colon, ascending/descending colon) , sacrum/coccyx: Rectal, S-colon/rectum, anus) are learned in advance, and the digestive system regions of the thoracic, lumbar, and sacral/coccyx sections are outputted based on this prior learning.

10 is a view showing a state in which the digestive organs and bleeding regions are marked on an axial plane image before and after injection of a contrast medium.

Referring to Figure 10, in the bleeding region of the digestive tract in the left lumbar region, (a) before contrast medium injection, (b) 20 seconds after contrast medium injection, (c) 60 seconds after contrast medium injection, The bleeding location/shape display unit 170 displays the digestive organs and bleeding regions on the axial plane image as shown in the figure on the right.

11 is a diagram illustrating a state in which the digestive organs and bleeding regions are marked according to bitwise AND and bitwise OR in axial plane images before and after injection of a contrast medium.

As shown in FIG. 11, axial images according to 'Bitwise AND (bitwise OR) and 'Bitwise OR (bitwise OR)' for each input image 20 seconds after contrast medium injection and 60 seconds after contrast medium injection The digestive tract and bleeding area are marked respectively.

12 is a diagram illustrating a state in which positions are marked on axial, coronal, and sagittal images according to a bleeding detection result of an image after injection of a contrast medium.

As shown in FIG. 12 , for each input image 20 seconds and 60 seconds after the injection of the contrast agent, the locations are displayed on the axial, coronal, and sagittal images according to the bleeding detection results, respectively.

14 is a diagram illustrating the calculation of the amount of bleeding based on an image after injection of a contrast medium.

Referring to FIG. 14 , (a) is an image 20 seconds after contrast medium injection, and (b) is an image 20 seconds after contrast medium injection, respectively, based on the 3D projection direction image. It is a conceptual representation of output.

As described above, the active bleeding detection system and method using abdominal CT image analysis according to the present invention detects the exact location of bleeding in the CT image of a bleeding patient inside the digestive tract, thereby directly / There is an effect that can help indirectly.

In addition, there is an advantage that the diagnosis can be made without delay for bleeding patients at all hospitals and clinics where radiologists do not reside.

In addition, it has the advantage of being able to represent the shape of the bleeding in 3D and to calculate the amount of bleeding and the bleeding rate.

In addition, when bleeding occurs in multiple locations, there is an advantage in that the priority required for emergency treatment can be designated and indicated by calculating the amount of bleeding or the bleeding rate.

As mentioned above, although the present invention has been described in detail through preferred embodiments, the present invention is not limited thereto, and it is common in the art that various changes and applications can be made without departing from the technical spirit of the present invention. self-explanatory to the technician. Accordingly, the true protection scope of the present invention should be construed by the following claims, and all technical ideas within the equivalent scope should be construed as being included in the scope of the present invention.

100: (Invention) Active bleeding detection system using abdominal CT image analysis
110: image input unit 120: image alignment unit
130: image segmentation unit 130a: spinal segment segmentation model
130b: digestive organ segmentation model 140: bleeding detection unit
150: bleeding shape component 160: first aid ranking unit
170: bleeding position / shape display unit 180: control unit
190: communication department

Claims (19)

an image input unit for receiving an abdominal CT image from the outside;
an image alignment unit for aligning images before and after injection of a contrast medium in the CT image received through the image input unit;
In the aligned image, the vertebral part is divided into a plurality of sections using the spine section segmentation model, and the digestive system is divided into a plurality of sections by using the digestive organ segmentation model in correspondence with the plurality of sections of the divided spine. with installments;
a bleeding detection unit for examining bleeding in the digestive tract divided into the plurality of sections and determining whether bleeding is detected according to the bleeding test;
a bleeding shape configuration unit that configures a shape of the detected bleeding when the bleeding is detected in the determination of the bleeding detection unit;
an emergency treatment priority designation unit for designating a bleeding emergency treatment priority when bleeding is detected at a plurality of locations in the determination of the bleeding detection unit;
a bleeding location/shape display unit that visualizes the location and shape of the bleeding detected in the plurality of locations and displays it on a screen; and
By controlling the status check and operation of the image input unit, image alignment unit, image segmentation unit, bleeding detection unit, bleeding shape configuration unit, first aid ranking unit, and bleeding location/shape display unit, and determining whether the analysis of the CT image is finished, and a controller for returning the analysis process to the step of receiving an image by the image input unit if it is not completed, and terminating the analysis process if the end result is an active bleeding detection system using abdominal CT image analysis.
According to claim 1,
The CT image aligned by the image alignment unit is a pre-phase image, an arterial-phase image after contrast medium injection, and a second set time after contrast medium injection (Portal-phase) An active bleeding detection system using an abdominal CT image analysis including images and images after a third set time has elapsed (delayed-phase) after injection of a contrast medium.
According to claim 1,
Active bleeding detection system using abdominal CT image analysis in which the image segmentation unit divides the spine into a plurality of sections, dividing the spine into three sections: thoracic, lumbar, and sacral/coccyx.
According to claim 1,
When the image segmentation unit divides the digestive system into a plurality of sections using the digestive organ segmentation model, the digestive organ segmentation model corresponds to the segmented vertebrae segment, a thoracic segment segmentation model, a lumbar segment segmentation model, and a sacral/caudal segment Active bleeding detection system using abdominal CT image analysis with segmentation model.
5. The method of claim 4,
When the image segmentation unit divides the digestive system into a plurality of sections using the digestive system segmentation model, the digestive system is divided into the esophagus, stomach, ascending/descending colon, and transverse colon segment using the thoracic segment segmentation model, the Active bleeding detection system using abdominal CT image analysis that divides each into a transverse colon and ascending/descending colon section using a lumbar segment segmentation model, and into a rectal and s-colon segment using the sacral/caudal segment segmentation model.
According to claim 1,
When the bleeding detection unit determines whether bleeding is detected according to the bleeding test of the digestive tract, abdominal CT image analysis is performed to determine whether bleeding is detected by comparing the images before and after the contrast medium injection using the digestive tract bleeding detection model. Active Bleeding Detection System.
7. The method of claim 6,
The digestive system bleeding detection model is an active bleeding detection system using abdominal CT image analysis including a thoracic vertebrae bleeding detection model, a lumbar vertebrae bleeding detection model, and a sacral/caudal region bleeding detection model.
According to claim 1,
Active bleeding detection system using abdominal CT image analysis to configure the bleeding shape in 3D when the bleeding shape component configures the bleeding shape.
According to claim 1,
When bleeding is detected at multiple locations, the first aid priority designation unit assigns a bleeding first aid priority when a plurality of bleeding is found in images 20 seconds, 60 seconds, and 100 seconds or more after contrast medium injection. Active bleeding detection system using abdominal CT image analysis to prioritize first aid treatment by calculating the amount of bleeding or bleeding rate at the found site.
According to claim 1,
Active bleeding detection system using abdominal CT image analysis further comprising a communication unit for transmitting/receiving data or information with an external device.
Activity based on an active bleeding detection system using abdominal CT image analysis including an image input unit, an image alignment unit, an image segmentation unit, a bleeding detection unit, a bleeding shape configuration unit, a first aid ranking unit, a bleeding location/shape display unit, and a control unit A method for detecting bleeding comprising:
a) the image input unit receiving an abdominal CT image from the outside;
b) aligning, by the image aligning unit, images before and after injection of a contrast medium in the CT image received through the image input unit;
c) The image segmentation unit divides the vertebral part into a plurality of sections using the vertebral section segmentation model in the aligned image, and divides the digestive system into a plurality of sections using the digestive system segmentation model in correspondence with a plurality of sections of the divided spine dividing each into
d) examining, by the bleeding detection unit, bleeding in the digestive system divided into the plurality of sections;
e) determining, by the bleeding detection unit, whether bleeding is detected according to the bleeding test of the digestive organs;
f) if bleeding is detected in the determination of step e), configuring the bleeding shape by the bleeding shape component;
g) if bleeding is detected in the determination of step e), determining by the bleeding detection unit whether bleeding is detected at a plurality of locations;
h) if bleeding is detected in a plurality of locations in the determination of step g), designating the bleeding first aid priority by the first aid priority designation unit;
i) the bleeding location/shape display unit visualizes the location and shape of the bleeding detected in the plurality of locations and displays it on a screen; and
j) determining whether the control unit ends the analysis of the CT image, returning the analysis process to step a) if it is not finished, and terminating the analysis process if it is finished. .
12. The method of claim 11,
In step a), the abdominal CT image is a pre-phase image, an arterial-phase image after contrast medium injection, and a second set time after contrast medium injection (Portal-phase) A method for detecting active bleeding using an abdominal CT image analysis including an image and an image after a third set time has elapsed (delayed-phase) after injection of a contrast medium.
12. The method of claim 11,
In step c), when the image segmentation unit divides the spine into a plurality of sections, the method for detecting active bleeding using abdominal CT image analysis divides the spine into three sections: thoracic, lumbar, and sacral/coccyx.
12. The method of claim 11,
In step c), the digestive system segmentation model corresponds to the segmented vertebral segment, a thoracic segment segmentation model, a lumbar segment segmentation model, and a sacral/caudal segment segmentation model. A method for detecting active bleeding using abdominal CT image analysis.
15. The method of claim 14,
When the image segmentation unit divides the digestive system into a plurality of sections using the digestive system segmentation model, the digestive system is divided into the esophagus, stomach, ascending/descending colon, and transverse colon segment using the thoracic segment segmentation model, the A method for detecting active bleeding using abdominal CT image analysis that divides the transverse colon and ascending/descending colon section using the lumbar segment segmentation model, and the rectum and s-colon segment using the sacral/caudal segment segmentation model, respectively. 12. The method of claim 11,
In step e), when the bleeding detection unit determines whether bleeding is detected according to the bleeding test of the digestive tract, it is determined whether bleeding is detected by comparing the images before and after the contrast medium injection using the digestive tract bleeding detection model. A method for detecting active bleeding using abdominal CT image analysis.
17. The method of claim 16,
The digestive system bleeding detection model is an active bleeding detection method using abdominal CT image analysis, including a thoracic vertebrae bleeding detection model, a lumbar vertebral region bleeding detection model, and a sacral/caudal region bleeding detection model.
12. The method of claim 11,
In step f), when the bleeding shape component constitutes the bleeding shape, the active bleeding detection method using abdominal CT image analysis to configure the bleeding shape in 3D.
12. The method of claim 11,
In step h), when the first aid ranking unit detects bleeding in multiple locations, in assigning the bleeding first aid priority, multiple bleeding is detected in the images 20 seconds, 60 seconds, and 100 seconds or more after contrast medium injection. If found, a method for detecting active bleeding using abdominal CT image analysis to prioritize first aid by calculating the amount of bleeding or bleeding rate at the site where bleeding was found.

Images