WO2023163287A1

WO2023163287A1 - Method and apparatus for analyzing medical image

Info

Publication number: WO2023163287A1
Application number: PCT/KR2022/008379
Authority: WO
Inventors: 기리시스리니바산; 김한석; 유영성; 피재우; 아킬라페루말라; 아터테이쇼크
Original assignee: 주식회사 피맥스
Priority date: 2022-02-25
Filing date: 2022-06-14
Publication date: 2023-08-31
Also published as: KR102525396B1

Abstract

The present disclosure relates to a method and apparatus for analyzing a medical image, wherein a method according to an embodiment of the present disclosure may comprise the steps of: acquiring a 3D medical image generated by photographing the body of a subject; by dividing the medical image on the basis of an anatomical direction corresponding to each of at least one pre-trained analysis model, generating at least one 2D cross-sectional image corresponding to each of the at least one analysis model; generating image information in which a plurality of body component regions are divided by inputting the 2D cross-sectional image to the corresponding analysis model; and from the image information, calculating information about at least one of a proportion and a volume of the plurality of body components.

Description

Medical image analysis method and apparatus

The technical idea of the present disclosure relates to a method and apparatus for analyzing a medical image.

Until now, the body composition analysis method has mainly consisted of measuring fat and muscle mass based on bioelectric impedance analysis (BIA). This analyzes body composition using the principle that when a fine current is applied to the body, a relatively large current flows through the body's water and muscle, and a relatively small current flows through the skeleton and fat.

Even in such a conventional technique, analysis results of body composition for each part can be obtained, but this is based on a two-dimensional analysis and is limited to a large range such as arms, upper body, lower body, and legs. That is, there is a problem in that the resolution of body composition analysis is low, so that the distribution of body components according to parts cannot be accurately known.

In addition, conventional methods relying on electrical signals rather than images have low accuracy due to high dependence on the analysis environment. That is, the electrical signal varies depending on the posture of the examinee or the amount of moisture in the body. For example, errors may also occur due to factors such as edema, excessive water intake, or a sudden change in posture before examination. In addition, in the method of measuring the microcurrent through the skin, errors also occur depending on the condition of the examinee's skin (eg, calluses).

In addition, since the conventional method provides only relatively simple information (e.g., muscle mass versus fat mass, etc.), it is impossible to analyze quantitative data for predicting a disease caused by an imbalance in body composition for each body part of a patient or diagnosing a disease. It also has limitations.

Therefore, there is a need for a new analysis technique capable of accurately analyzing body composition for various body parts of an examinee.

A technical problem to be achieved by a method and apparatus for analyzing medical images according to the technical idea of the present disclosure is medical image analysis capable of providing more accurate information about body components included in the whole body or body part of an examinee based on a 3D medical image. It is to provide a method and an apparatus therefor.

Technical tasks to be achieved by the medical image analysis method and apparatus according to the technical idea of the present disclosure are not limited to the tasks mentioned above, and other tasks not mentioned will be clearly understood by those skilled in the art from the description below.

According to one aspect of the technical concept of the present disclosure, a medical image analysis method includes obtaining a 3D medical image generated by photographing a body of an examinee; generating at least one 2D cross-sectional image corresponding to each of the at least one analysis model by segmenting the medical image based on an anatomical direction corresponding to each of the at least one analysis model that has been learned; generating image information obtained by dividing a plurality of body component regions by inputting the 2D cross-sectional image to the corresponding analysis model; and calculating information on at least one of ratios and volumes of a plurality of body components from the image information.

According to an exemplary embodiment, the medical image may include at least one of a 3D computed tomography (3D-CT) image and a magnetic resonance image (MRI).

According to an exemplary embodiment, the medical image may be a T1-weighted magnetic resonance image.

According to an exemplary embodiment, the 2D cross-sectional image may include at least one of an axial image, a sagittal image, and a coronal image.

According to an exemplary embodiment, the analysis model includes a first analysis model and a second analysis model, and the first analysis model determines the body composition through a plurality of cross-sectional images generated from a plurality of three-dimensional training medical images. The second analysis model may be pre-trained to divide the body component region through at least one of a plurality of sagittal plane images and coronal plane images generated from the training medical image.

According to an exemplary embodiment, the body component may include muscle, visceral adipose tissue (VAT), subcutaneous adipose tissue (SAT), intramuscular adipose tissue (IMAT), skeleton, and at least one may include organs of

According to an exemplary embodiment, the calculating of information on at least one of a ratio and a volume of the body component may include calculating the number of pixels for each body component region from the image information generated corresponding to each of the 2D sectional images. doing; obtaining information on an actually measured size of a pixel from photographing information of the medical image; and calculating the volume of each body component based on the number of pixels of the body component region and information about the measured size.

According to an exemplary embodiment, the method further includes dividing at least one body part to be analyzed from the medical image, and generating the 2D cross-sectional image is performed on the divided body part from the medical image. It can be.

According to an exemplary embodiment, the method may further include generating body component comparison information for each body part by comparing information on at least one of a ratio and a volume of the body components calculated from the different body parts.

According to one aspect of the technical concept of the present disclosure, a medical image analysis apparatus includes at least one processor; a memory for storing a program executable by the processor; and the processor, by executing the program, obtains a 3D medical image generated by photographing the examinee's body, and generates the medical image based on an anatomical direction corresponding to each of at least one pre-learned analysis model. By dividing, at least one 2D sectional image corresponding to each of the analysis models is generated, and image information obtained by dividing a plurality of body component regions is generated by inputting the 2D sectional image to the analysis model corresponding to the analysis model. Information on at least one of ratios and volumes of a plurality of body components may be calculated from the information.

According to the medical image analysis method and apparatus according to embodiments according to the technical idea of the present disclosure, by providing quantified body composition analysis from medical images, it is not dependent only on human visual judgment, thereby reducing the fatigue of diagnosis, Variation due to subjective judgment for each user can also be reduced.

The effects that can be obtained by the medical image analysis method and apparatus according to the technical idea of the present disclosure are not limited to the above-mentioned effects, and other effects not mentioned are common knowledge in the art to which the present disclosure belongs from the description below. will be clearly understandable to those who have

A brief description of each figure is provided in order to more fully understand the figures cited in this disclosure.

1 is a flowchart illustrating a method for analyzing a medical image according to an embodiment of the present disclosure.

FIG. 2 is a flowchart for explaining an embodiment of step S140 of FIG. 1 .

3 and 4 are diagrams exemplarily illustrating body fat analysis information calculated by a medical image analysis method according to an embodiment of the present disclosure.

5 is a flowchart illustrating a method for analyzing a medical image according to an embodiment of the present disclosure.

6 and 7 are diagrams exemplarily illustrating body fat analysis information calculated by a medical image analysis method according to an embodiment of the present disclosure.

8 is a block diagram briefly illustrating the configuration of a medical image analysis apparatus according to an embodiment of the present disclosure.

Since the technical spirit of the present disclosure may be subject to various changes and may have various embodiments, specific embodiments are illustrated in the drawings and described in detail. However, this is not intended to limit the technical spirit of the present disclosure to specific embodiments, and should be understood to include all changes, equivalents, or substitutes included in the scope of the technical spirit of the present disclosure.

In describing the technical idea of the present disclosure, if it is determined that a detailed description of a related known technology may unnecessarily obscure the subject matter of the present disclosure, the detailed description will be omitted. In addition, numbers (eg, first, second, etc.) used in the description process of the present disclosure are only identifiers for distinguishing one component from another component.

In addition, in the present disclosure, when one component is referred to as "connected" or "connected" to another component, the one component may be directly connected or directly connected to the other component, but in particular Unless otherwise described, it should be understood that they may be connected or connected via another component in the middle.

In addition, terms such as "~ unit", "~ group", "~ character", and "~ module" described in the present disclosure mean a unit that processes at least one function or operation, which includes a processor, a micro Processor (Micro Processor), Micro Controller, CPU (Central Processing Unit), GPU (Graphics Processing Unit), APU (Accelerate Processor Unit), DSP (Digital Signal Processor), ASIC (Application Specific Integrated Circuit), FPGA (Field Programmable Gate Array) may be implemented by hardware or software or a combination of hardware and software.

In addition, it is intended to make it clear that the classification of components in the present disclosure is merely a classification for each main function in charge of each component. That is, two or more components to be described below may be combined into one component, or one component may be divided into two or more for each more subdivided function. In addition, each component to be described below may additionally perform some or all of the functions of other components in addition to its main function, and some of the main functions of each component may be performed by other components. Of course, it may be dedicated and performed by .

The method according to an embodiment of the present disclosure may be performed in a personal computer having computing capability, a workstation, a computer device for a server, or a separate device for this purpose.

Also, the method may be performed on one or more computing devices. For example, at least one or more steps of the method 100 according to an embodiment of the present disclosure may be performed in a client device and other steps may be performed in a server device. In this case, the client device and the server device may be connected through a network to transmit and receive calculation results. Alternatively, method 100 may be performed by distributed computing technology.

Hereinafter, embodiments of the present disclosure will be described in detail in turn.

1 is a flowchart for explaining a method for analyzing a medical image according to an embodiment of the present disclosure, and FIG. 2 is a flowchart for explaining an embodiment of step S140 of FIG. 1 .

In step S110, the device may obtain a 3D medical image generated by photographing the examinee's body.

For example, the medical image may be received from an external database server or may be captured by a photographing device connected to the device through wired or wireless communication.

In the embodiment, the medical image may be an image generated by photographing the entire body of the examinee (ie, the entire body), but is not limited thereto, and according to an embodiment, an image generated by photographing a part of the examinee's body may be used. can

In an embodiment, the medical image may include at least one of a 3D computed tomography (3D-CT) image and a magnetic resonance image (MRI). In particular, in the case of a magnetic resonance image, a T1-weighted magnetic resonance image may be applied. In the case of a T1-weighted magnetic resonance image, since the contrast value difference between fat and muscle is high, segmentation accuracy of body component regions can be further improved through the analysis model described in detail below.

Meanwhile, in an embodiment, the method 100 may further include pre-processing the medical image after step S110. For example, in this step, the device may first filter whether the medical image is a normal image that can be analyzed through a preprocessing algorithm, and convert the data format to be suitable for the input format of the analysis model. In addition, the device may correct a contrast value deviation due to an error or noise of a photographing device through an image normalization process.

In step S120, the apparatus may generate at least one 2D sectional image corresponding to each of the at least one analysis model by segmenting the 3D medical image based on the anatomical direction corresponding to each of the at least one analysis model that has been learned in advance. . Here, the 2D cross-sectional image may include at least one of an axial image, a sagittal image, and a coronal image.

In an embodiment, the device cuts at least a part of the examinee's body included in the 3D medical image perpendicularly to the first direction at regular intervals while moving in a first direction (eg, a direction from the head to the legs). By doing so, at least one 2-dimensional cross-sectional image (eg, cross-sectional image) corresponding to the first analysis model may be generated. In addition, in the embodiment, the device moves at least a part of the examinee's body included in the 3D medical image in a second direction (eg, from the left arm to the right arm), and at regular intervals perpendicular to the second direction. By cutting, at least one 2-dimensional cross-sectional image (eg, sagittal image) corresponding to the second analysis model may be generated. In addition, in the embodiment, the device cuts at least a part of the examinee's body included in the 3D medical image perpendicularly to the third direction at regular intervals while moving in a third direction (eg, from the front to the rear direction). By doing so, at least one 2-dimensional cross-sectional image (eg, coronal image) corresponding to the third analysis model may be generated. Here, the first direction, the second direction, and the third direction may be directions perpendicular to each other.

In an embodiment, a plurality of analysis models may be configured, and each may be pre-learned to segment a body component region based on different 2D cross-sectional images. For example, the analysis model includes a first analysis model and a second analysis model, and the first analysis model is pre-trained to segment body component regions through a plurality of cross-sectional images generated from a plurality of three-dimensional training medical images. The second analysis model may be pre-trained to segment the body component region through at least one of a plurality of sagittal plane images and coronal plane images generated from the training medical images.

In an embodiment, the device may select at least one analysis model to be used for analysis from among a plurality of analysis models. For example, the device selects an analysis model based on a user input, or when the user selects all or part of a body region included in a medical image as an analysis target, an analysis model optimized for analysis of the corresponding region is automatically selected. can be configured to

In operation S130 , image information obtained by dividing a plurality of body component regions may be generated by inputting the generated 2D cross-sectional image to a corresponding analysis model. Here, the body composition may include muscle, visceral adipose tissue (VAT), subcutaneous adipose tissue (SAT), intramuscular adipose tissue (IMAT), skeleton, and at least one organ. .

Each analysis model may include at least one network function. That is, prior learning may be performed so that the network function divides (or detects) body component regions from medical images and/or cross-sectional images generated therefrom in advance through learning data.

A network function may be used in the same sense as a neural network and/or a neural network. A neural network (neural network) may be composed of a set of interconnected computational units, which may be generally referred to as nodes, and these nodes may be referred to as neurons. A neural network generally includes a plurality of nodes, and the nodes constituting the neural network may be interconnected by one or more links. In this case, some of the nodes constituting the neural network may configure one layer based on distances from the first input node. For example, a set of nodes having a distance of n from the first input node may constitute n layers. The neural network may include a deep neural network (DNN) including a plurality of hidden layers in addition to an input layer and an output layer.

In an embodiment, the image information may be in the form of emphasizing and displaying each divided body component region with a different color for each 2D cross-sectional image, and the generated image information is separately provided to the user according to a user input. It can be.

In step S140, the device may calculate information on at least one of ratios and volumes of a plurality of body components from the image information.

For example, the device may extract the area of each body component region from image information, calculate the volume of each body component by combining them, and generate information about the ratio of body components according to the volume ratio between them.

In an embodiment, step S140 may include steps S141 to S143.

In step S141, the device may calculate the number of pixels for each body component region from image information generated corresponding to each 2D cross-sectional image.

Subsequently, in step S142, the device may obtain information on the actually measured size (horizontal and vertical length) of the pixel from photographing information (eg, Dicom tag information) of the medical image.

Next, in step S143, the device may calculate the volume of each body component based on the information about the number of pixels and the measured size of each body component region.

For example, the device may calculate the area of a body component included in each 2D cross-sectional image by multiplying the number of pixels of each body component region by the measured size. In addition, the device may calculate the volume occupied by each body component in the entire or part of the examinee's body based on the arrangement and division intervals between the plurality of consecutive 2D sections.

Meanwhile, although not shown, the method 100 may further include providing information on the calculated body composition (ie, volume and/or ratio of the body composition) to the user. In this case, information on the body composition may be provided along with a 3D medical image, a body image of the examinee generated by reconstructing the image, and/or at least one 2D sectional image. Various methods may be applied to the provided image so that the user can visually identify the region occupied by each body component.

First, referring to FIG. 3 , one or more sagittal plane images and/or coronal plane images of the entire body are generated from a 3D medical image, and based on this, information on the ratio of each body component in the entire body of the examinee is calculated. can Here, information on the ratio of body components may include a ratio occupied by the corresponding body component in the whole and/or a relative ratio between two or more body components. The calculated body composition ratio may be provided to the user (or user terminal) in the form of a predetermined table.

In this case, each body component region segmented (or detected) by the analysis model may be displayed on a sectional image (ie, a sagittal image and/or a coronal image) in a predetermined manner. For example, when a user selects a specific body composition, an area corresponding to the corresponding body composition may be displayed in a predetermined color.

Referring to FIG. 4 , one or more cross-sectional images of the entire body may be generated from a 3D medical image, and information on the volume and/or ratio of each body component in the entire body of the examinee may be calculated based on the generated cross-sectional images. Similarly, the volume and/or ratio of the calculated body composition may be provided to the user (or user terminal) in a table format.

Similar to FIG. 3 , each body component region segmented by the analysis model may be displayed on a sectional image (ie, a cross-sectional image) in a predetermined manner so as to be identified separately.

Depending on the embodiment, the user may change the cross-sectional image in which the body composition region is displayed through a predetermined input. For example, a shape corresponding to the entire body of the examinee may be displayed, and when a user selects a specific location thereof, a body component region may be displayed on a cross-sectional image corresponding to the location.

In an embodiment, information on body composition may be provided separately for each body part. For example, when a predetermined interface is provided to a user and the user selects a specific body part (eg, abdomen) based on this, the device displays image information generated from a 2D sectional image corresponding to the body part. It may be configured to analyze and calculate body fat information and provide it to the user.

Here, redundant descriptions of steps similar to those of the method 100 described above with reference to FIG. 1 in the method 500 of FIG. 5 will be omitted, and the description of FIG. 1 will be referred to.

When a 3D medical image of the examinee is acquired in step S510, in step S520, the apparatus may segment at least one body part to be analyzed from the medical image.

In an embodiment, step S520 may be performed based on user input. That is, when a user sets a specific body part as an analysis target through a predetermined interface, the device extracts the body part set by the user from the medical image and divides it from other body parts. For example, the user may set one or more body parts, such as legs, arms, and abdomen, as analysis targets.

In an embodiment, step S520 may be performed by a pretrained network function.

In step S530, the device may generate one or more 2D cross-sectional images of the corresponding body part by segmenting the body part set as the analysis target in an anatomical direction corresponding to each analysis model.

In steps S540 and S550, the device generates image information obtained by dividing a plurality of body component regions by inputting the two-dimensional cross-sectional image into a corresponding analysis model, and from the image information, the ratio and volume of the plurality of body components included in each body part. Information on at least one of them can be calculated. In an embodiment, step S550 may be performed in the same manner as step S140 described above with reference to FIG. 2 .

In step S560, the device may generate body composition comparison information for each body part. That is, when a plurality of body parts are set as analysis targets by the user, the device may generate comparison information by comparing information such as the body component volume and/or ratio of each body part with each other.

The generated comparison information may be provided to a user or the like in a predetermined manner.

First, referring to FIG. 6 , when the abdomen is set as an analysis target, an abdominal region may be segmented from a medical image. Subsequently, information on the ratio of each body component included in the abdomen of the examinee may be calculated based on the plurality of cross sections generated from the abdominal region. The calculated body composition ratio may be provided to the user (or user terminal) in the form of a predetermined table.

In this case, each body component region segmented by the analysis model may be displayed in a predetermined manner so as to be distinguished and identified on a cross-sectional image (ie, a cross-sectional image) or a 3D image of an abdominal region separated from a medical image.

Referring to FIG. 7 , when both arms are set as analysis targets, regions corresponding to both arms may be segmented from a medical image. Subsequently, information on the ratio of each body component included in the examinee's left arm and right arm may be calculated based on a plurality of cross sections respectively generated from both arms. Subsequently, comparison information for comparing the body composition ratios of both arms is generated, and such comparison information may be provided to a user (or user terminal) in a predetermined table format.

Each body component region segmented by the analysis model may be displayed in a predetermined manner in a 3D image of each arm separated from a sectional image or a medical image to be identified separately.

The communication unit 810 may receive input data (medical images, etc.) for performing medical image analysis. The communication unit 810 may include a wired/wireless communication unit. When the communication unit 810 includes a wired communication unit, the communication unit 810 may include a local area network (LAN), a wide area network (WAN), a value added network (VAN), and a mobile communication network ( Mobile Radio Communication Network), a satellite communication network, and one or more components that enable communication through a mutual combination thereof. In addition, when the communication unit 810 includes a wireless communication unit, the communication unit 810 transmits and receives data or signals wirelessly using cellular communication, a wireless LAN (eg, Wi-Fi), and the like. can In an embodiment, the communication unit may transmit/receive data or signals with an external device or an external server under the control of the processor 840 .

The input unit 820 may receive various user commands through external manipulation. To this end, the input unit 820 may include or connect one or more input devices. For example, the input unit 820 may be connected to various input interfaces such as a keypad and a mouse to receive user commands. To this end, the input unit 820 may include an interface such as a thunderbolt as well as a USB port. In addition, the input unit 820 may receive an external user command by including or combining various input devices such as a touch screen and buttons.

The memory 830 may store programs and/or program commands for operation of the processor 840 and may temporarily or permanently store input/output data. The memory 830 is a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (eg SD or XD memory, etc.), RAM , SRAM, ROM (ROM), EEPROM, PROM, magnetic memory, a magnetic disk, it may include at least one type of storage medium.

In addition, the memory 830 may store various network functions and algorithms, and may store various data, programs (one or more instructions), applications, software, commands, codes, etc. for driving and controlling the device 700. there is.

The processor 840 may control the overall operation of the device 800 . Processor 840 may execute one or more programs stored in memory 830 . The processor 840 may mean a central processing unit (CPU), a graphics processing unit (GPU), or a dedicated processor on which methods according to the technical idea of the present disclosure are performed.

In an embodiment, the processor 840 obtains a 3D medical image generated by photographing the examinee's body, and divides the medical image based on an anatomical direction corresponding to each of at least one pre-learned analysis model. , At least one 2D cross-sectional image corresponding to each of the analysis models is generated, and image information obtained by dividing a plurality of body component regions is generated by inputting the 2-dimensional cross-sectional image to the corresponding analysis model, and from the image information Information on at least one of ratios and volumes of a plurality of body components may be calculated.

In an embodiment, the processor 840 calculates the number of pixels for each of the body component regions from the image information generated corresponding to each of the 2D cross-sectional images, and information about the actually measured size of a pixel from the photographing information of the medical image. is obtained, and the volume of each body component can be calculated based on the number of pixels of the body component region and information on the measured size.

In an embodiment, the processor 840 may divide at least one body part to be analyzed from the medical image, and generate the 2D cross-sectional image of the divided body part from the medical image.

In an embodiment, the processor 840 may generate body component comparison information for each body part by comparing information on at least one of a ratio and a volume of the body components calculated from the different body parts.

The method according to an embodiment of the present disclosure may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer readable medium. The computer readable medium may include program instructions, data files, data structures, etc. alone or in combination. Program instructions recorded on the medium may be those specially designed and configured for the purpose of the present disclosure, or may be known and usable to those skilled in computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic media such as floptical disks. - includes hardware devices specially configured to store and execute program instructions, such as magneto-optical media, and ROM, RAM, flash memory, and the like. Examples of program instructions include high-level language codes that can be executed by a computer using an interpreter, as well as machine language codes such as those produced by a compiler.

In addition, the method according to the disclosed embodiments may be provided by being included in a computer program product. Computer program products may be traded between sellers and buyers as commodities.

A computer program product may include a S/W program and a computer-readable storage medium in which the S/W program is stored. For example, a computer program product may include a product in the form of a S/W program (eg, a downloadable app) that is distributed electronically through a manufacturer of an electronic device or an electronic marketplace (eg, Google Play Store, App Store). there is. For electronic distribution, at least a part of the S/W program may be stored in a storage medium or temporarily generated. In this case, the storage medium may be a storage medium of a manufacturer's server, an electronic market server, or a relay server temporarily storing SW programs.

A computer program product may include a storage medium of a server or a storage medium of a client device in a system composed of a server and a client device. Alternatively, if there is a third device (eg, a smart phone) that is communicatively connected to the server or the client device, the computer program product may include a storage medium of the third device. Alternatively, the computer program product may include a S/W program itself transmitted from the server to the client device or the third device or from the third device to the client device.

In this case, one of the server, the client device and the third device may execute the computer program product to perform the method according to the disclosed embodiments. Alternatively, two or more of the server, the client device, and the third device may execute the computer program product to implement the method according to the disclosed embodiments in a distributed manner.

For example, a server (eg, a cloud server or an artificial intelligence server) may execute a computer program product stored in the server to control a client device communicatively connected to the server to perform a method according to the disclosed embodiments.

Although the embodiments have been described in detail above, the scope of the present disclosure is not limited thereto, and various modifications and improvements of those skilled in the art using the basic concepts of the present disclosure defined in the following claims are also within the scope of the present disclosure. belongs to

Claims

In the medical image analysis method,

obtaining a 3D medical image generated by photographing the examinee's body;

generating at least one 2D cross-sectional image corresponding to each of the at least one analysis model by segmenting the medical image based on an anatomical direction corresponding to each of the at least one analysis model that has been learned;

generating image information obtained by dividing a plurality of body component regions by inputting the 2D cross-sectional image to the corresponding analysis model; and

And calculating information on at least one of ratios and volumes of a plurality of body components from the image information.
According to claim 1,

The method of claim 1 , wherein the medical image includes at least one of a three-dimensional computed tomography (3D-CT) image and a magnetic resonance image (MRI).
According to claim 1,

The method of claim 1 , wherein the medical image is a T1 weighted magnetic resonance image.
According to claim 1,

The 2D cross-sectional image includes at least one of an axial image, a sagittal image, and a coronal image.
According to claim 1,

The analysis model includes a first analysis model and a second analysis model,

The first analysis model is pretrained to segment the body component region through a plurality of cross-sectional images generated from a plurality of three-dimensional training medical images;

The second analysis model is pre-learned to segment the body component region through at least one of a plurality of sagittal plane images and coronal plane images generated from the training medical images.
According to claim 1,

The body composition includes muscle, visceral adipose tissue (VAT), subcutaneous adipose tissue (SAT), intramuscular adipose tissue (IMAT), skeleton, and at least one organ. .
According to claim 1,

Calculating information on at least one of the ratio and volume of the body composition,

calculating the number of pixels for each of the body component regions from the image information generated corresponding to each of the 2D sectional images;

obtaining information on an actually measured size of a pixel from photographing information of the medical image; and

The method further comprises calculating a volume of each of the body components based on the number of pixels of the body component region and the information about the measured size.
According to claim 1,

Segmenting at least one body part to be analyzed from the medical image;

The generating of the 2D cross-sectional image is performed on the segmented body part in the medical image.
According to claim 8,

The method further comprising generating body component comparison information for each body part by comparing information on at least one of a ratio and a volume of the body components calculated from the different body parts.
In the medical image analysis device,

at least one processor;

a memory for storing a program executable by the processor; and

The processor, by executing the program, acquires a three-dimensional medical image generated by photographing the examinee's body, and divides the medical image based on an anatomical direction corresponding to each of at least one pre-learned analysis model. By doing so, at least one 2D sectional image corresponding to each of the analysis models is generated, and image information obtained by dividing a plurality of body component regions is generated by inputting the 2D sectional image to the corresponding analysis model, and the image information An apparatus for calculating information on at least one of ratios and volumes of a plurality of body components from
A computer program stored in a recording medium to execute the method of any one of claims 1 to 9.