KR102032611B1 - Method and application for determining cardiovascular disease using ct image - Google Patents

Abstract

A method of analyzing a medical image by a computer, the method comprising: acquiring a CT image of a heart of an object and inputting the acquired CT image to a trained first model to obtain information about the inside of a coronary artery of the object A method is disclosed, comprising a step.

Description

Methods and applications for determining cardiovascular lesions using CT images {METHOD AND APPLICATION FOR DETERMINING CARDIOVASCULAR DISEASE USING CT IMAGE}

The present invention relates to methods and applications for determining cardiovascular lesions using CT images.

As a method for determining cardiovascular lesions, various methods such as electrocardiogram, exercise load test, CT scan, angiography, intravascular ultrasonography, and near infrared spectroscopy are used.

CT imaging is a computed tomography (CT) scan using a CT scanner. X-rays or ultrasounds are projected onto the human body from various angles and then reconstructed by a computer to process images of the internal cross-section of the human body as images.

Angiography and intravascular ultrasonography, on the other hand, are invasive test methods that are difficult for patients to use.

Recently, development of a method of acquiring necessary information by processing various medical images using deep learning has been actively conducted.

Deep learning is defined as a set of machine learning algorithms that attempts to achieve high levels of abstraction (summarizing key content or functions in large amounts of data or complex data) through a combination of several nonlinear transformations. Deep learning can be seen as a field of machine learning that teaches computers how people think in a large framework.

An object of the present invention is to provide a method and application for determining a cardiovascular lesion using a CT image.

Problems to be solved by the present invention are not limited to the above-mentioned problems, and other problems not mentioned will be clearly understood by those skilled in the art from the following description.

According to an aspect of the present invention, a method for analyzing a medical image by a computer includes: obtaining a CT image of a heart of an object and inputting the acquired CT image to a trained first model Acquiring information about the inside of the coronary artery of the subject.

In addition, acquiring information about the inside of the coronary artery of the subject may include acquiring information about a plaque inside the coronary artery of the subject.

In addition, the learned first model, CT images of the heart of the subject, IntraVascular UltraSound (IVUS) of the subject and Near InfraRed Spectrocopy (NIRS) results in the coronary vessels of the subject It may be characterized as a model trained using.

In addition, acquiring information about the coronary artery of the subject may include acquiring information about the internal coronary artery of the subject.

In addition, the trained first model may be a model trained using a CT image of the heart of the subject and a coronary vascular ultrasound image (IVUS) of the subject.

In addition, acquiring information about the inside of the coronary arteries may further include obtaining information on the deposits inside the coronary artery by inputting the acquired information on the coronary internal form into a trained second model. It may include.

In addition, the trained second model may be a model trained using a coronary vascular ultrasound image of the subject (IVUS) and a near infrared spectroscopy (NIRS) result of the coronary vessel of the subject.

The acquiring of information on the inside of the coronary artery may further include determining a section causing angina in the coronary artery of the subject.

In addition, the trained first model may be a model trained using a CT image of a heart of a subject and a myocardial fraction reserve (FFR) test result in a coronary vessel of the subject. have.

An application stored in a computer-readable recording medium is provided to perform the method according to an aspect of the present invention for solving the above problems.

Other specific details of the invention are included in the detailed description and drawings.

According to the disclosed embodiment, there is an effect of replacing the invasive inspection method through CT imaging, which is a non-invasive inspection method.

In addition, as well as the morphological significance of the cardiovascular information as well as symptomatic significance can be obtained according to the disclosed embodiment, there is an effect that can more accurately determine the lesions of the patient, and specifically determine the treatment method.

Effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the following description.

1 is a conceptual diagram illustrating a system for determining a cardiovascular state of a subject using a CT image, according to an exemplary embodiment.
2 is a flowchart schematically illustrating a method of determining a cardiovascular state of a subject using a CT image, according to an exemplary embodiment.
3 is a flowchart illustrating a method of determining a cardiovascular state of a subject using a CT image according to an exemplary embodiment.
4 is a diagram illustrating examples of cardiac CT images, coronary IVUS images, and coronary TVC images of a subject.
5 is a flowchart illustrating a method of determining a heart lesion of a subject, according to an exemplary embodiment.
6 is a diagram illustrating an example of obtaining a coronary TVC image from a heart CT image.
7 is a diagram illustrating an example of a method for training a model capable of acquiring TVC images from IVUS images of coronary arteries.
8 is a diagram illustrating an example of a method of segmenting an intravascular ultrasound image.
9 is a flowchart illustrating a method of acquiring a coronary artery section causing angina from a heart CT image according to an embodiment.
10 is a flowchart illustrating a method of determining a heart disease of a subject according to an exemplary embodiment.

Advantages and features of the present invention and methods for achieving them will be apparent with reference to the embodiments described below in detail with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but can be embodied in various different forms, and the present embodiments only make the disclosure of the present invention complete, and those of ordinary skill in the art to which the present invention belongs. It is provided to fully inform the skilled worker of the scope of the invention, which is defined only by the scope of the claims.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. In this specification, the singular also includes the plural unless specifically stated otherwise in the phrase. As used herein, "comprises" and / or "comprising" does not exclude the presence or addition of one or more other components in addition to the mentioned components. Like reference numerals refer to like elements throughout, and "and / or" includes each and all combinations of one or more of the mentioned components. Although "first", "second", etc. are used to describe various components, these components are of course not limited by these terms. These terms are only used to distinguish one component from another. Therefore, of course, the first component mentioned below may be a second component within the technical spirit of the present invention.

Unless otherwise defined, all terms used in the present specification (including technical and scientific terms) may be used in a sense that can be commonly understood by those skilled in the art. In addition, terms that are defined in a commonly used dictionary are not ideally or excessively interpreted unless they are specifically defined clearly.

As used herein, the term "part" or "module" refers to a hardware component such as software, FPGA, or ASIC, and the "part" or "module" plays certain roles. However, "part" or "module" is not meant to be limited to software or hardware. The “unit” or “module” may be configured to be in an addressable storage medium or may be configured to play one or more processors. Thus, as an example, a "part" or "module" may include components such as software components, object-oriented software components, class components, and task components, processes, functions, properties, Procedures, subroutines, segments of program code, drivers, firmware, microcode, circuitry, data, databases, data structures, tables, arrays, and variables. Functions provided within components and "parts" or "modules" may be combined into smaller numbers of components and "parts" or "modules" or into additional components and "parts" or "modules". Can be further separated.

The spatially relative terms " below ", " beneath ", " lower ", " above ", " upper " It can be used to easily describe a component's correlation with other components. Spatially relative terms are to be understood as including terms in different directions of components in use or operation in addition to the directions shown in the figures. For example, when flipping a component shown in the drawing, a component described as "below" or "beneath" of another component may be placed "above" the other component. Can be. Thus, the exemplary term "below" can encompass both an orientation of above and below. Components may be oriented in other directions as well, so spatially relative terms may be interpreted according to orientation.

As used herein, "image" may mean multi-dimensional data composed of discrete image elements (eg, pixels in a 2D image and voxels in a 3D image). For example, the image may include a medical image of the object obtained by the CT imaging apparatus.

In the present specification, a “CT (Computed Tomography) image” may mean a composite image of a plurality of X-ray images obtained by photographing an object while rotating about at least one axis of the object.

As used herein, an "object" may be a person or an animal, or part or all of a person or an animal. For example, the subject may include at least one of organs such as the liver, heart, uterus, brain, breast, abdomen, and blood vessels. In addition, the "object" may be a phantom. Phantom means a material having a volume very close to the density and effective atomic number of an organism, and may include a sphere phantom having properties similar to the body.

As used herein, a "user" may be a doctor, a nurse, a clinical pathologist, a medical imaging professional, or the like, and may be a technician who repairs a medical device, but is not limited thereto.

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

1 is a conceptual diagram illustrating a system for determining a cardiovascular state of a subject using a CT image, according to an exemplary embodiment.

The system shown in FIG. 1 includes a CT system 100 and a computer 200.

Since the CT system 100 may provide a cross-sectional image of the object 10, an internal structure of the object (for example, an organ such as a kidney or a lung, etc.) does not overlap with a general X-ray imaging device. There is this.

The CT system 100 may provide, for example, a relatively accurate cross-sectional image of an object by acquiring and processing image data having a thickness of 2 mm or less tens or hundreds per second. Conventionally, there has been a problem that only the horizontal cross section of the object is expressed, but it is overcome by the appearance of various image reconstruction techniques as follows. Three-dimensional reconstruction imaging techniques include the following techniques.

SSD (Shade surface display): A technique for displaying only voxels having a certain HU value as an initial three-dimensional imaging technique.

Maximum intensity projection (MIP) / minimum intensity projection (MinIP): A 3D technique showing only those having the highest or lowest HU value among the voxels constituting the image.

VR (volume rendering): A technique that can adjust the color and transmittance of the voxels constituting the image for each region of interest.

Virtual endoscopy: A technique that allows endoscopic observation in three-dimensional images reconstructed by the VR or SSD technique.

MPR (multi planar reformation): An image technique for reconstructing different cross-sectional images. It is possible to reconstruct freedom in the direction desired by the user.

Editing: Various techniques for arranging surrounding voxels to make it easier to observe the point of interest in VR.

VOI (voxel of interest): A technique for representing only a selected area in VR.

The CT system 100 according to an embodiment of the present invention may include various types of devices, and the structure of the CT system 100 shown in FIG. 1 is described as an example.

Referring to FIG. 1, the CT system 100 may include a gantry 102, a table 105, an X-ray generator 106, and an X-ray detector 108.

The gantry 102 may include an X-ray generator 106 and an X-ray detector 108.

The object 10 may be located on the table 105.

The table 105 may move in a predetermined direction (eg, at least one of up, down, left, and right) during the CT imaging process. In addition, the table 105 may be tilted or rotated by a predetermined angle in a predetermined direction.

In addition, the gantry 102 may also be inclined by a predetermined angle in a predetermined direction.

In one embodiment, the CT system 100 includes a workstation 110 capable of acquiring and analyzing a photographed CT image.

The workstation 110 transmits the acquired CT image to the computer 200, and the computer 200 determines a cardiovascular state of the object using the CT image according to the disclosed embodiment.

2 is a flowchart schematically illustrating a method of determining a cardiovascular state of a subject using a CT image, according to an exemplary embodiment.

The method shown in FIG. 2 comprises steps that are processed in time series in the computer 200 shown in FIG.

In step S210, the first model is learned. Training of the first model may be performed by the computer 200 or may be performed by at least one other computer.

The trained first model may be stored in the computer 200 or may be stored in a server and used by the computer 200.

In one embodiment, the first model is trained based on machine learning. In one embodiment, the first model is trained based on deep learning.

Deep learning is defined as a set of machine learning algorithms that attempts to achieve high levels of abstraction (summarizing key content or functions in large amounts of data or complex data) through a combination of several nonlinear transformations. Deep learning can be seen as a field of machine learning that teaches computers how people think in a large framework.

When there is any data, it is represented in a form that can be understood by a computer (for example, in the case of an image, the pixel information is represented by a column vector), and a lot of research (how to do it better) is applied to learning. How to make representational techniques and how to build models for learning them. As a result of these efforts, various deep learning techniques have been developed. Deep learning techniques include Deep Neural Networks (DNN), Convolutional Deep Neural Networks (CNN), Reccurent Neural Networks (RNN), and Deep Belief Networks (DBN). have.

Deep Neural Networks (DNNs) are artificial neural networks (ANNs) made up of a plurality of hidden layers between an input layer and an output layer.

The above-described algorithm or learning method is described as an example, and the first model according to the disclosed embodiment may be learned by any learning method. For example, the first model according to the disclosed embodiment includes supervised learning in which input data and output data are provided as learning data, unsupervised learning in which no separate labeled data is provided, and learning results. Feedback may be learned by reinforcement learning, but is not limited thereto.

In one embodiment, the first model is a CT image of the heart of the subject, IntraVascular UltraSound (IVUS) of the subject and Near InfraRed Spectrocopy (NIRS) results in the coronary vessels of the subject It is learned using at least one.

The first model is trained using at least a portion of the CT image of the heart of the object as input data and at least one of an IVUS image and a NIRS result of the object as output data.

In an embodiment, the first model is trained using at least a portion of a CT image of the heart of the object as input data and a TVC image of the object as output data.

In the disclosed embodiment, the TVC image is acquired by Infraredx's TVC imaging system, and provides information about the size and chemical composition of the lesion through ultrasound imaging and near infrared spectroscopy of blood vessels.

In the disclosed embodiment, the TVC image is used as a term for an image obtained through ultrasound imaging and near infrared spectroscopy of a blood vessel, and is not necessarily limited to an image acquired by an Infraredix TVC imaging system.

In operation S220, the computer 200 acquires a heart CT image of the object.

In one embodiment, the computer 200 acquires at least some of the heart CT images of the subject.

In one embodiment, the computer 200 acquires a part of the coronary artery from the heart CT image of the subject.

In operation S230, the computer 200 obtains information about the inside of the coronary artery of the object by inputting the CT image acquired in operation S220 to the first model learned in operation S210.

In one embodiment, the information about the interior of the coronary artery of the subject includes information about the plaque inside the coronary artery of the subject.

In one embodiment, the computer 200 obtains information about deposits inside the coronary artery of the subject by inputting the CT image acquired in step S220 to the first model learned in step S210.

For example, the computer 200 inputs at least a portion of the CT image acquired in step S220 (eg, a portion of the coronary artery) to the first model trained in step S210 to provide an image of the inside of the coronary artery of the subject. Obtain NIRS results.

For example, the computer 200 inputs at least a portion of the CT image acquired in step S220 (eg, a portion of the coronary artery) to the first model trained in step S210 to provide an image of the inside of the coronary artery of the subject. Acquire TVC images.

In one embodiment, the information about the internal coronary artery of the subject includes information about the shape of the internal coronary artery of the subject.

In one embodiment, the computer 200 obtains information about the shape of the inside of the coronary artery of the subject by inputting the CT image acquired in step S220 to the first model learned in step S210.

For example, the computer 200 inputs at least a portion (eg, a portion of the coronary artery) of the CT image acquired in step S220 into the first model learned in step S210 to coronary internal ultrasound image of the object. Acquire (IVUS image).

3 is a flowchart illustrating a method of determining a cardiovascular state of a subject using a CT image according to an exemplary embodiment.

The method shown in FIG. 3 comprises steps that are processed in time series in the computer 200 shown in FIG.

In addition, the method illustrated in FIG. 3 is an embodiment of the method illustrated in FIG. 2, and although the description omitted with reference to FIG. 3 applies to the description described with reference to FIG. 2.

In step S210, the first model and the second model are learned.

The first model corresponds to the first model described with respect to FIG. 2.

In one embodiment, the first model is trained using at least a portion of a CT image of the heart of the object as input data and an IVUS image of the object as output data.

For example, the first model is trained using a portion of the CT image of the heart of the subject as the input data and an IVUS image of the coronary artery as the output data.

In one embodiment, the second model is trained using ultrasound images in the coronary vessels of the subject and near infrared spectroscopy (NIRS) results in the coronary vessels of the subject.

For example, the second model is trained using the ultrasound image in the coronary vessel of the subject as input data and the NIR result in the coronary vessel of the subject as the output data.

For example, the second model is trained by using an ultrasound image in the coronary vessel of the subject as input data and a TVC image in the coronary vessel of the subject as output data.

In operation S320, the computer 200 acquires a heart CT image of the object.

In operation S330, the computer 200 obtains information about the internal coronary artery shape of the object by inputting the heart CT image of the object to the first model learned in operation S310.

In one embodiment, the computer 200 obtains a coronary blood vessel ultrasound image of the subject by inputting the heart CT image of the subject to the first model learned in operation S310.

In operation S340, the computer 200 obtains information about deposits in the coronary vessels of the subject by inputting the coronary vessel intravascular ultrasound image of the subject acquired in operation S330 to the second model learned in operation S310.

In one embodiment, the computer 200 obtains information about the NIRS result inside the coronary vessel of the subject by inputting the coronary vessel intravascular ultrasound image of the subject acquired in operation S330 to the second model learned in operation S310. .

In one embodiment, the computer 200 obtains a TVC image inside the coronary vessel of the subject by inputting the coronary vessel intravascular ultrasound image of the subject acquired in operation S330 to the second model learned in operation S310.

Referring to FIG. 4, an example of a heart CT image 400, a coronary IVUS image 410, and a coronary TVC image 420 of a subject is shown.

The computer 200 learns a portion of the subject's heart CT image 400 (eg, a portion of the coronary artery 402) by using the method described with reference to FIGS. 2 and 3. Input to the model to obtain a coronary IVUS image 410 of the subject or a coronary TVC image 420 of the subject.

The computer 200 inputs the acquired coronary IVUS image 410 of the object into the trained model to obtain a coronary TVC image 420 of the object.

5 is a flowchart illustrating a method of determining a heart lesion of a subject, according to an exemplary embodiment.

In general, the method of determining the heart lesion of the subject in the initial stage is an electrocardiography (ECG) shown in step S510. There are many different types of ECG tests, including 24-hour ECG and exercise-loaded ECG tests.

In step S520, a heart CT scan is performed to observe the external shape of the heart. If it is difficult to determine the lesion through the heart CT scan, the heart lesion is observed through angiography shown in step S530.

Further, in step S540, the intravascular ultrasound (IVUS) image is taken to acquire the internal shape of the blood vessel. However, since it is difficult to determine the trait of each component constituting the internal form of the blood vessel only with the IVUS image, in step S550, a TVC image is obtained by utilizing near infrared spectroscopy.

Each diagnostic method illustrated in FIG. 5 does not have to be performed sequentially or in each step, and may be selectively used according to the opinion of a specialist.

However, steps S530 to S550 correspond to invasive inspections, and are difficult to use because they are relatively expensive inspection methods.

According to the method described with reference to FIGS. 2 to 4, an IVUS image or a TVC image of the coronary artery may be acquired using the heart CT image of the object acquired in step S520.

According to step S522, the computer 200 may obtain a coronary TVC image of the subject from the heart CT image of the subject using the trained model.

According to step S524, the computer 200 acquires a coronary IVUS image of the subject from the heart CT image of the subject using the trained model, and inputs the acquired IVUS image to the learned model to obtain a coronary TVC image of the subject. Can be obtained.

In an embodiment, the learned model corresponding to step S522 may be a form in which the learned models corresponding to step S524 and step S542 are merged.

In an embodiment, the trained model corresponding to step S522 may be a trained model using a heart CT image and a coronary TVC image.

6 is a diagram illustrating an example of obtaining a coronary TVC image from a heart CT image.

Referring to FIG. 6, a heart CT image 600 of an object is illustrated.

The TVC image is obtained using coronary internal ultrasound images and near infrared spectroscopy, and using TVC, cross-sectional views 650 and 660 from different sides of the coronary arteries can be obtained as shown in FIG. 6.

Referring to the cross-sectional view 650 shown in FIG. 6, the location of the vessel wall 652 of the coronary artery and the deposit 654 in the vessel wall 652 is shown. In particular, deposits 654 represent deposits of dangerous traits that can cause angina in the vessel wall 652.

In addition, another type of cross-sectional view 660 can be obtained by using TVC. Referring to cross-sectional view 660, along with an IVUS image corresponding to one cross section in the coronary artery, information about the side on which the deposit of the dangerous trait is formed is displayed at the edge of the IVUS image.

In one embodiment, the computer 200 may acquire a TVC image 650 or 660 using the coronary artery portion 602 of the heart CT image 600.

As one step for obtaining a TVC image 650 or 660 of the coronary artery from the cardiac CT image 600, a model trained to obtain a TVC image of the coronary artery from an IVUS image of the coronary artery may be utilized. .

Referring to FIG. 7, an example of a method for training a model capable of acquiring TVC images from IVUS images of coronary arteries is shown.

7 shows a TVC image 700 of a coronary artery of a subject. The TVC image 700 may be divided into a portion 704 representing an internal vascular form that can be obtained using IVUS and a portion 702 representing a trait of an intravascular deposit that can be obtained using NIRS.

In an embodiment, the computer 200 may cut the TVC image 700 at a predetermined angle. For example, the computer 200 may cut the TVC image 700 into rectangles 710 and 720 inclined by 10 degrees, as shown in FIG. 7.

The computer 200 uses a portion 702 representing the trait of intravascular deposits, contained in each cut rectangle, to determine whether the vascular internal shape is normal or abnormal, included in each cut rectangle. You can judge.

For example, each of the rectangles 710 and 720 cut as shown in FIG. 7 may be processed into training data 730 and 740 labeled normal and abnormal, respectively.

The computer 200 uses the labeled training data 730 and 740 to train the model.

When a new IVUS image is acquired, the computer 200 cuts the acquired IVUS image at a predetermined angle and inputs each cut image to the learned model to determine whether it is normal or abnormal.

The computer 200 may restore the TVC image from the IVUS image based on the determination result of each cropped image.

8 is a diagram illustrating an example of a method of segmenting an intravascular ultrasound image.

Referring to FIG. 8, an intravascular ultrasound image 800 is shown.

In training the model using the intravascular ultrasound image 800 as input data or output data, each part of the image 800 may be divided or labeled.

As shown in FIG. 8, the image 800 includes an image of a blood vessel tissue 810, a deposit 820, and an internal blood vessel 830.

The image 800 is automatically or manually labeled or labeled with the vascular tissue 810, the deposit 820, and the blood vessel interior 830.

In particular, the deposit 820 or the vessel interior 830 may be marked with a region of interest, and may be utilized to train the model so that the model can distinguish the deposit or vessel interior from an intravascular ultrasound image. .

9 is a flowchart illustrating a method of acquiring a coronary artery section causing angina from a heart CT image according to an embodiment.

The method shown in FIG. 9 comprises steps that are processed in time series in the computer 200 shown in FIG.

The Myocardial Fractional Flow Reserve (FFR) test measures pressure in the coronary arteries to determine which sections in the coronary arteries actually cause ischemia, i.e., angina.

Blood vessels that appear to be narrowed or blocked may not actually cause ischemia, and may also develop ischemia even if there are no symptoms or if the exercise test is negative.

The narrow blood vessel does not necessarily need to perform a stent procedure, and in consideration of the side effects, it may be desirable to perform the procedure only on the site where the tongue blood is actually generated. Therefore, in addition to the morphological observation described with reference to FIGS. 1 to 8, a method for observing symptomatic significance will be described in FIGS. 9 and 10.

In step S910, the third model is trained. In one embodiment, the third model is trained using a CT image of the heart of the subject and an FFR test result of the subject.

In an embodiment, the third model may be trained using at least a portion of a CT image of the heart of the object as input data and an FFR test result of the object as output data.

In operation S920, the computer 200 acquires a heart CT image of the object.

In operation S930, the computer 200 may input a cardiac CT image of the object acquired in operation S920, to the third model learned in operation S910, to determine an interval causing angina in the coronary artery of the object.

In one embodiment, the computer 200 inputs the heart CT image of the object acquired in step S920 to the third model learned in step S910 to obtain FFR data about the coronary artery of the object.

As shown in FIG. 10, a CT image of the heart of the subject is taken in step S1010, and if necessary, morphological information inside the coronary artery is acquired as described with reference to FIGS. 1 to 8.

In step S1020, the FFR test is performed to determine the portion of the subject's coronary artery ischemia.

In an embodiment, the computer 200 may acquire information about a part in which ischemia occurs in the coronary artery of the subject using the heart CT image of the subject acquired in operation S1010 and the learned model (operation S1015). .

The steps of a method or algorithm described in connection with an embodiment of the present invention may be implemented directly in hardware, in a software module executed by hardware, or by a combination thereof. Software modules may include random access memory (RAM), read only memory (ROM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, hard disk, removable disk, CD-ROM, or It may reside in any form of computer readable recording medium well known in the art.

In the above, embodiments of the present invention have been described with reference to the accompanying drawings, but those skilled in the art to which the present invention pertains may implement the present invention in other specific forms without changing the technical spirit or essential features. I can understand that. Therefore, it should be understood that the embodiments described above are exemplary in all respects and not restrictive.

10: object
100: CT system
102: gantry
105: table
106: X-ray generator
108: X-ray detector
110: workstation
200: computer

Claims (10)

In the cardiovascular lesion determination method performed on a computer,
An image acquiring step of acquiring a first image of the cardiovascular body of the object; And
And outputting a second image by inputting the first image to a first model so as to identify traits of blood vessel deposits from the first image.
The second image includes a diagnostic auxiliary region indicating whether blood vessels are normal,
The second image corresponds to an anatomical area provided by the first image as a sum of a plurality of first rectangles.
The plurality of first rectangular regions are generated and disposed in a form rotated at a predetermined angle at a position corresponding to the center of the first image, and are generated based on at least a portion of the region of the first image.
An area corresponding to the first rectangular area of the diagnostic auxiliary area is included in at least a part of the first rectangular area;
An area corresponding to the first rectangular area among the diagnostic auxiliary areas indicates whether a blood vessel included in the first rectangular area is normal.
How to determine cardiovascular lesions.
According to claim 1,
The first model is trained through a learning set consisting of images obtained by the same photographing method as the second image,
Inputting an area corresponding to the first image in the image of the learning set to the first model, and outputting an image including the diagnostic auxiliary area in the image of the learning set,
How to determine cardiovascular lesions. According to claim 1,
The cardiovascular lesion determination method further includes a preprocessing step of dividing the first image into a plurality of second rectangular regions,
The plurality of second rectangular areas are areas divided at predetermined angles at positions corresponding to the center of the first image based on the first image.
How to determine cardiovascular lesions. According to claim 1,
The first image includes an ultrasound image of cardiovascular of the subject,
The second image includes a near infrared spectral spectrum result of the cardiovascular of the subject together with the first image.
How to determine cardiovascular lesions. According to claim 1,
The diagnostic auxiliary region indicates a normal state of blood vessels included in the first rectangular region as a near infrared spectral spectrum.
How to determine cardiovascular lesions. The method of claim 5,
When it is determined that the blood vessel is abnormal when the region represented by the first color included in the diagnostic assistance region is equal to or greater than a predetermined ratio among the diagnostic assistance regions,
How to determine cardiovascular lesions. According to claim 1,
The first image is an image that is output when a third image, which is a heart CT image of the object, is input to the second model.
The second model is trained through a learning set composed of images obtained by the same photographing method as the third image.
How to determine cardiovascular lesions. The method of claim 7, wherein
The second image is output through the first model and the second model from the third image,
The method may further include determining an interval causing angina from the second image in the heart of the object included in the third image.
How to determine cardiovascular lesions. The method of claim 8,
The section causing the angina is determined based on the diagnostic assistance region,
How to determine cardiovascular lesions. An application stored on a computer-readable recording medium for performing the method of any one of claims 1 to 9.

Images