KR102243830B1 - System for providing integrated medical diagnostic service and method thereof - Google Patents

Abstract

The present invention relates to a system for providing an integrated medical diagnosis service and a method thereof. In the method of providing an integrated medical diagnosis service performed by an integrated medical diagnosis service providing system that provides diagnosis information through medical image analysis, a) an informal Collecting and storing raw medical image data, and providing standardized medical image data through pre-processing of the raw medical image data; b) learning to extract features of an anatomical structure or lesion of the body from the medical image data to provide a classification model, and extracting features of newly input medical image data based on the classification model; c) providing diagnostic information by deriving a lesion for a corresponding medical image using the extracted features; d) inferring at least one suspicious lesion area that is a major factor in deriving the lesion in the medical image based on the diagnosis information, and then providing visual information of the lesion area using the inferred suspicious lesion area; And e) generating at least one keyword matching the anatomy of the body for the lesion area using the visual information of the lesion area, and then generating a diagnostic report using the diagnosis information, visual information of the lesion area, and keywords. It is to include the step of providing.

Description

System for providing integrated medical diagnosis service and method thereof {SYSTEM FOR PROVIDING INTEGRATED MEDICAL DIAGNOSTIC SERVICE AND METHOD THEREOF}

The present invention relates to a system and method for providing an integrated medical diagnosis service capable of performing an accurate diagnosis on a medical image using a deep learning algorithm and providing a visual and verbal medical explanation for the diagnosis.

In medical care, the importance of optimal diagnosis or prescription through data analysis is increasing. However, with the development of various medical devices and hospital information systems, the amount of medical data generated and stored is explosively increasing.According to IBM's estimate, more than 200 times the amount of data available to medical staff is expected to be collected in 2020. have. In contrast to this rapid increase in data, the shortage of medical staff is expected to accelerate.

In Korea, there are only 10 emergency reading specialists who professionally perform emergency reading during the late-night time, accounting for only 0.27% of the 3,700 radiology specialists. Therefore, there is a risk that the time and accuracy of diagnosis, such as stroke or cardiovascular disease, are fatal to the prognosis, as the reading is delayed until the time the specialist reads it, or the time the specialist is going to work.

In addition, it takes more than 10 years to train a corresponding specialist in the field of radiology specializing in medical image reading, and even if there is an environment in which medical staff can read medical images, the match rate for reading among the same specialists is 60. Since it is only at the level of %~65%, there is a problem that the reading results of medical images are inconsistent according to the knowledge and clinical experience of the specialist.

As such, there is a need for a solution to the low consistency of judgment due to the uncertainty and ambiguity of medical data.

Recently, artificial intelligence, especially deep learning technology, has been actively introduced into the medical field, which is rapidly developing accuracy and efficiency in various fields such as image, voice, and natural language recognition. Research on increasingly diverse and complex medical data analysis and diagnostic assistance solutions with strengths of prediction consistency of deep learning, scalability that is not limited by time and space, and accuracy comparable to human accuracy. Development is actively taking place.

In the case of deep learning-based medical data analysis, the most research results in medical image analysis have been published so far, and results that match or exceed doctors' diagnosis and reading ability in the detection and quantification of lesions, and classification of lesions. Is appearing.

However, since deep learning-based medical data analysis technology self-learns from a vast amount of medical data through a deep learning network and makes classification decisions, humans cannot know the internal process of the deep learning network, and thus, it is large in the judgment of the diagnoser. Reliability can be compromised in the medical field where responsibility is involved.

In order to solve this problem, research is being actively conducted on a method to analyze the operating principle in a deep learning network and a method to visualize the area that has a major influence on the classification result. However, the deep learning-based medical image diagnosis technology does not have an intuitive description of the output because it is through a complex operation on input data up to the output.

Meanwhile, in the existing deep learning-based medical data analysis and diagnosis technology, the method of measuring medical images varies depending on the patient's condition, and the data types, storage methods, and analysis methods are very different from hospital to hospital. Due to the lack of consistency, it is practically impossible to pre-process data in an automated batch method. Therefore, regardless of the format of medical image data, there is a problem in that it is not possible to categorize medical images and provide explanations capable of verbal interpretation at a level that can be understood by patients for patients and experts.

Republic of Korea Patent Registration No. 10-1623431 (Name of invention: pathology diagnosis classification device for medical images and pathology diagnosis system using the same)

Korean Patent Registration No. 10-1857624 (Name of invention: medical diagnosis method reflecting clinical information and device using the same)

In order to solve the above-described problem, the present invention performs accurate medical diagnosis using deep learning algorithms for various medical images according to an embodiment of the present invention, and provides visual and verbal diagnosis at a level that can be understood by a user. The purpose is to provide.

However, the technical problem to be achieved by the present embodiment is not limited to the technical problem as described above, and other technical problems may exist.

As a technical means for achieving the above technical problem, a method for providing an integrated medical diagnosis service according to an embodiment of the present invention includes an integrated medical diagnosis performed by an integrated medical diagnosis service providing system that provides diagnosis information through medical image analysis. A service providing method comprising the steps of: a) collecting and storing unstructured raw medical image data, and providing standardized medical image data through pre-processing of the raw medical image data; b) learning to extract features of an anatomical structure or lesion of the body from the medical image data to provide a classification model, and extracting features of newly input medical image data based on the classification model; c) providing diagnostic information by deriving a lesion for a corresponding medical image using the extracted features; d) inferring at least one suspicious lesion area that is a major factor in deriving the lesion in the medical image based on the diagnosis information, and then providing visual information of the lesion area using the inferred suspicious lesion area; And e) generating at least one keyword matching the anatomy of the body for the lesion area using the visual information of the lesion area, and then generating a diagnostic report using the diagnosis information, visual information of the lesion area, and keywords. It is to include the step of providing.

In an integrated medical diagnosis service providing system according to an embodiment of the present invention, there is provided an integrated medical diagnosis service providing system for providing diagnosis information through medical image analysis, comprising: a communication unit that processes transmission and reception of signals related to medical images through a communication network; A memory in which a program for performing an integrated medical diagnosis service providing method based on the medical image analysis is recorded; And a control unit for executing the program, wherein the control unit collects and stores unstructured raw medical image data by execution of the program, and the standardized medical image through preprocessing of the raw medical image data Provides data, learns to extract features of an anatomical structure or lesion of the body from the medical image data to provide a classification model, and extracts features of newly input medical image data based on the classification model, Deriving a lesion for a corresponding medical image using the extracted features to provide diagnostic information, and inferring at least one suspicious lesion area that is a major factor in deriving the lesion in the medical image based on the diagnostic information Then, the visual information of the lesion area is provided using the inferred suspected lesion area, and at least one keyword matching the anatomy of the body for the lesion area is generated using the visual information of the lesion area, and the diagnosis It generates and provides a diagnostic report using information, visual information of the lesion area, and keywords.

According to the above-described problem solving means of the present invention, in order to automatically provide diagnosis results for various medical images regardless of the format of the medical image and at the same time, visualize and visualize areas that are major factors in diagnosis to explain the basis of the diagnosis. By generating keywords for specific body parts in the designated area, it is possible to provide a diagnostic report capable of explaining visual and verbal diagnosis at a level that can be understood by a doctor, patient, or guardian.

In the present invention, the contribution to the output can be visualized using a deep learning model structure that maintains spatial information, and the basis for diagnosis can be presented. Reliability can be increased.

In addition, the present invention provides a medical image captured with a medical device to an integrated medical diagnosis service providing system without a patient or guardian directly visiting a professional medical staff or hospital, thereby enabling the analysis of the medical image, thereby providing a diagnosis probability and a basis for diagnosis. The presented medical findings can be checked, and even professional medical staff can verify their diagnosis results through the diagnosis report through the integrated medical diagnosis service providing system, thereby preventing misdiagnosis.

The present invention can be used not only as a diagnostic tool for medical images and an auxiliary tool for diagnosis explanation, but also as a medical auxiliary device that automatically refers to past medical images and makes additional prescriptions possible according to the progress of a disease.

1 is a diagram showing the configuration of a system for providing an integrated medical diagnosis service according to an embodiment of the present invention.
2 is a flowchart illustrating a method of providing an integrated medical diagnosis service according to an embodiment of the present invention.
3 is a diagram illustrating a classification process of a medical image based on a classification model according to an embodiment of the present invention.
4 is a diagram illustrating an example of diagnostic information, visual information of a lesion area, and text information through MRI image analysis according to an embodiment of the present invention.
5 is a flowchart illustrating an image segmentation and analysis process according to an embodiment of the present invention.
6 is an exemplary diagram of a deep learning model based on an MR image by a method for providing an integrated medical diagnosis service according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those of ordinary skill in the art can easily implement the present invention. However, the present invention may be implemented in various different forms and is not limited to the embodiments described herein. In the drawings, parts irrelevant to the description are omitted in order to clearly describe the present invention, and similar reference numerals are attached to similar parts throughout the specification.

Throughout the specification, when a part is said to be "connected" with another part, this includes not only "directly connected" but also "electrically connected" with another element interposed therebetween. . In addition, when a part "includes" a certain component, it means that other components may be further included, and one or more other features, not excluding other components, unless specifically stated to the contrary. It is to be understood that it does not preclude the presence or addition of any number, step, action, component, part, or combination thereof.

The following examples are detailed descriptions for aiding understanding of the present invention, and do not limit the scope of the present invention. Accordingly, the invention of the same scope performing the same function as the present invention will also belong to the scope of the present invention.

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

1 is a diagram showing the configuration of a system for providing an integrated medical diagnosis service according to an embodiment of the present invention.

Referring to FIG. 1, the integrated medical diagnosis service providing system 100 includes a communication unit 110, a memory 120, a control unit 130, a display unit 140, and a database 150.

In detail, the communication unit 110 provides a communication interface necessary for providing a signal transmitted and received to the integrated medical diagnosis service providing system 100 in the form of packet data in connection with a communication network. In this case, the communication unit 110 may be a device including hardware and software necessary for transmitting and receiving a signal such as a control signal or a data signal through a wired or wireless connection with another network device.

This communication unit 110 is a PACS (Picture Archiving Communication System, medical image storage and transmission system) for storage and transmission of medical images generated through various diagnostic medical equipment, medical mobile added value to users (doctors, patients, guardians, etc.) A mobile diagnostic medical image integrated service system for providing a service (mobile value-added medical service), a user terminal, and the like are connected through a communication network to transmit and receive various medical data including medical images and diagnostic information.

In the memory 120, a program for performing a method of providing an integrated medical diagnosis service is recorded. In addition, the control unit 130 performs a function of temporarily or permanently storing data processed. Here, the memory 120 may include a volatile storage medium or a non-volatile storage medium, but the scope of the present invention is not limited thereto.

The controller 130 controls the entire process of providing a method of providing an integrated medical diagnosis service based on deep learning. Each step performed by the control unit 130 will be described later with reference to FIG. 2.

Here, the controller 130 may include all types of devices capable of processing data, such as a processor. Here, the'processor' may refer to a data processing device embedded in hardware having a circuit physically structured to perform a function represented by, for example, a code included in a program or an instruction. As an example of a data processing device built into the hardware as described above, a microprocessor, a central processing unit (CPU), a processor core, a multiprocessor, and an application-specific integrated (ASIC) circuit), a field programmable gate array (FPGA), and the like, but the scope of the present invention is not limited thereto.

The display unit 140 outputs visual information or a diagnosis report of a lesion area, which is a major factor for diagnosis, under the control of the controller 130.

The database 150 stores data accumulated while performing a method of providing an integrated medical diagnosis service. The database 140 may store user-specific medical image data, diagnosis information, visual information of a lesion area, and a diagnosis report.

2 is a flowchart illustrating a method of providing an integrated medical diagnosis service according to an embodiment of the present invention, and FIG. 3 is a diagram illustrating a classification process of a medical image based on a classification model according to an embodiment of the present invention. 4 is a diagram illustrating an example of diagnostic information, visual information of a lesion area, and text information through MRI image analysis according to an embodiment of the present invention.

Referring to FIG. 2, in the method of providing an integrated medical diagnosis service, unstructured raw medical image data is collected and stored in a database, and standardized medical image data is obtained through preprocessing of the raw medical image data stored in the database ( S10). The controller 130 collects various raw medical images from a user regardless of a specific disease or type, including X-ray, CT, and MRI, as well as endoscopic, ultrasound, and pathological images, and stores them in a database.

Since the raw medical image data stored in the database is unstructured, the controller 130 processes it into standardized medical image data through a pre-processing process. That is, since raw medical image data is high-dimensional and not normalized in nature and contains unnecessary information for diagnosis, the controller 130 performs pre-processing including image alignment and normalization of intensity between images on the raw medical image data. After that, it is processed into the form of input data of a deep learning algorithm that reduces the data size.

If the medical image data is a brain image, the controller 130 uses skull stripping to remove unnecessary skull parts in the brain image, and the anterior Commissure and Posterior Commissure (ACPC) for the brain image. ACPC alignment, which performs image alignment by adjusting the position around the junction surface, and normalization, which matches the intensity between images, are performed.

The controller 130 provides a classification model by learning to extract features of an anatomical structure or lesion of a body from medical image data using a deep learning algorithm based on Convolutional Neural Networks (CNN) (S20). The controller 130 trains a classification model using a 3D CNN to extract features of medical image data including 3D spatial information, and learns weights for classification of medical images.

The controller 130 increases the total number of data to learn about the medical image by using a method such as flip, brightness adjustment, contrast adjustment, cropping, and pixel unit shift. In this case, in the deep learning algorithm, learning is to statistically adjust the weight between the input data and the output data, and the connection link between nodes in the neural network structure includes the weight.

As shown in FIG. 3, each class (Class1, Class2, etc.) for all subject images (Subj#1 to Subj#n) through an automatic classifier learned using features extracted from preprocessed medical image data. After obtaining the probability value to be included in ), the image features are extracted using the probability values of the suspected lesion area, and these are repeatedly learned.

The controller 130 extracts features of the medical image data newly input through the classification model (S30), and calculates diagnosis information by deriving a lesion for the medical image using the extracted features (S40). In this case, the controller 130 may automatically and quickly derive a disease corresponding to the medical image by assigning a weight to each medical image.

At this time, the diagnosis information is one of lesion presence information indicating the presence or absence of a lesion for the medical image, diagnosis name information through pathology diagnosis of the lesion, diagnosis probability information including a matching probability value for the diagnosis name, and lesion detail information regarding the location or size of the lesion. It can be created including the above.

The controller 130 infers at least one suspicious lesion area that is a major factor in deriving a lesion in the medical image based on the diagnosis information, and then calculates visual information of the lesion area using the inferred suspicious lesion area (S50). .

That is, the controller 130 infers at least one suspicious lesion area, which is a major factor in deriving lesions, from the medical image data through backtracking of the process of calculating the diagnostic information, and compares the inferred suspicious area with the corresponding medical image. Visual information of the lesion area is generated by projecting it onto an image plane of the same dimension.

The controller 130 extracts at least one keyword matching the anatomy of the body for the lesion area by using visual information on the lesion area, and integrates the extracted keyword and diagnostic information to explain the medical diagnosis. Is calculated (S60).

As shown in FIG. 4, the controller 130 retrieves the MRI image 200, which is the standardized medical image data, and performs image analysis. As a result, the diagnosis information 210 including the diagnosis name and diagnosis probability, and the main lesion area are Visual information 220 and text information 230 of the lesion area are calculated.

At this time, the visual information 220 of the lesion area enables intuitive interpretation of the medical image so that it is possible to check which body part has greatly influenced the diagnosis. In addition, the text information 230 includes a chart number, name, etc. of a patient as an object of the MRI image, and includes a verbal diagnosis description at a level that the patient or guardian can understand.

The controller 130 generates a diagnostic report by synthesizing diagnostic information, visual information of the lesion area, and text information, and provides it to the user (S70).

The controller 130 performs an image segmentation and analysis process to generate visual information of the lesion area at a keyword level for a specific body part.

5 is a flowchart illustrating an image segmentation and analysis process according to an embodiment of the present invention.

Referring to FIG. 5, the controller 130 divides medical image data into a preset number of subject images according to the anatomy of the body (S710). At this time, the controller performs a labeling operation of allocating a label for each subject image using a human anatomy atlas (S720). In the case of an MRI image, the MRI image is divided into subject images for each brain organ, and then labeling is performed using Brain Atlas so that each subject image corresponds to which organ.

The controller 130 analyzes the characteristics of the anatomical structure of the divided subject image, detects at least one subject image having a suspicious lesion area, and then derives the main lesion area using the weight (S730, S740).

The control unit 130 extracts an anatomical name of a specific body part that is a major factor in diagnosis as a keyword by matching the label assigned to the subject image for the major lesion area (S750 and S760).

In this way, when the control unit 130 divides the medical image into meaningful regions, it is possible to clearly distinguish between internal organs and the boundary line of a tumor in a medical image such as CT or MRI taken with a tomography. By synthesizing the divided subject images, the structure inside the body can be realized in three dimensions, so that the lesion can be diagnosed more accurately.

6 is an exemplary diagram of a deep learning model based on an MR image by a method for providing an integrated medical diagnosis service according to an embodiment of the present invention.

Referring to FIG. 6, when a user selects to load an MR image, an MR image of a patient and patient information are displayed on the screen. Patient information includes ID, name, gender, date of birth, and treatment date.

When the user selects to analyze the MR image, the unstructured MR image is pre-processed, and feature extraction and diagnosis results are derived based on the deep learning model.

The integrated medical diagnosis service provision system visualizes the derived diagnosis results and significant significant lesion areas that are the main factors in the diagnosis results, and creates keywords for specific body parts for the visualized areas, so that doctors and patients can understand them. Can generate diagnostic reports.

Therefore, the patient does not need to directly visit and hear the opinions of medical staff in each medical field, and if the captured medical image is transmitted to the integrated medical diagnosis service providing system after taking an image through a medical device, a diagnosis report is provided. You can check it through. Doctors can also prevent false diagnosis by checking the diagnosis report through the integrated medical diagnosis service providing system and verifying their diagnosis results.

The method for providing an integrated medical diagnosis service according to an embodiment of the present invention described above may be implemented in the form of a recording medium including instructions executable by a computer, such as a program module executed by a computer. Such recording media include computer-readable media, and computer-readable media may be any available media that can be accessed by a computer, and include both volatile and nonvolatile media, removable and non-removable media. In addition, computer-readable media includes computer storage media, which are volatile and nonvolatile embodied in any method or technology for storage of information such as computer-readable instructions, data structures, program modules or other data. , Removable and non-removable media are included.

The above description of the present invention is for illustrative purposes only, and those of ordinary skill in the art to which the present invention pertains will be able to understand that other specific forms can be easily modified without changing the technical spirit or essential features of the present invention. will be. Therefore, it should be understood that the embodiments described above are illustrative in all respects and are not limiting. For example, each component described as a single type may be implemented in a distributed manner, and similarly, components described as being distributed may also be implemented in a combined form.

The scope of the present invention is indicated by the claims to be described later rather than the detailed description, and all changes or modified forms derived from the meaning and scope of the claims and their equivalent concepts should be construed as being included in the scope of the present invention. do.

100: Integrated medical diagnosis service provision system
110: communication unit 120: memory
130: control unit 140: display unit
150: database

Claims (18)

A method for providing an integrated medical diagnosis service performed by an integrated medical diagnosis service providing system that provides diagnosis information through medical image analysis, the method comprising:
a) collecting and storing unstructured raw medical image data, and providing standardized medical image data through pre-processing of the raw medical image data;
b) learning to extract features of an anatomical structure or lesion of the body from the medical image data to provide a classification model, and extracting features of newly input medical image data based on the classification model;
c) providing diagnostic information by deriving a lesion for a corresponding medical image using the extracted features;
d) inferring at least one suspicious lesion area that is a major factor in deriving the lesion in the medical image based on the diagnosis information, and then providing visual information of the lesion area using the inferred suspicious lesion area; And
e) After generating at least one keyword matching the anatomy of the body for the lesion region using the visual information of the lesion region, a diagnostic report is generated using the diagnostic information, visual information of the lesion region, and keywords. Provided, wherein the visual information includes images obtained by dividing the medical image data, and the keyword is a name of an anatomical structure of the body that matches an image including the lesion suspicious region among the images divided by the medical image data The method of providing integrated medical diagnosis service. The method of claim 1,
The step a),
Providing the medical image data by processing the raw medical image data in the form of input data of the classification model,
The method of providing an integrated medical diagnosis service, wherein the size of the data is reduced after performing pre-processing including image alignment and normalization of intensity between images of the medical image data. The method of claim 2,
When the medical image data is a brain image,
The step of removing unnecessary cranial regions in the brain image, and performing image alignment by adjusting the position of the brain image around the junction of the Anterior Commissure and Posterior Commissure (ACPC) How to provide medical diagnostic services. The method of claim 1,
The step b) is to learn a weight for classification of the medical image by using a deep learning algorithm based on Convolutional Neural Networks (CNN), an integrated medical diagnosis service providing method. The method of claim 4,
In step b), the classification model is trained using a 3D CNN to extract features of medical image data including 3D spatial information. The method of claim 1,
Step d),
Inferring at least one suspicious lesion area, which is a major factor in deriving the lesion, through trace back to the process of deriving the diagnostic information; And
And generating visual information of the lesion region by projecting the inferred lesion suspicious region onto an image plane having the same dimension as the corresponding medical image. The method of claim 6,
Step d),
When the classification model uses a deep learning algorithm based on Convolutional Neural Networks (CNN), a feature that is a major factor in deriving the lesion among the extracted features is obtained through weights, and the feature is inferred as the lesion suspicious region. That is, a method of providing integrated medical diagnosis services. The method of claim 1,
The step e),
Dividing the medical image data into a predetermined number of subject images according to the anatomy of the body based on artificial intelligence;
Performing a labeling operation of allocating a label for each subject image using a human anatomy atlas;
Analyzing characteristics of the anatomical structure of the segmented subject image, detecting at least one subject image having a suspicious lesion area, and then deriving a major lesion area using weights; And
And extracting an anatomical name for a body part as a keyword by matching a label assigned to the subject image corresponding to the derived major lesion area. The method of claim 1,
Step e),
By integrating the extracted keyword and diagnostic information to generate text information for explaining a medical diagnosis and include it in the diagnosis report. The method of claim 1,
The step c),
Lesion presence information indicating the presence or absence of a lesion in the medical image using the extracted features, diagnosis name information through pathology diagnosis of the lesion, diagnosis probability information including a coincidence probability value for the diagnosis name, lesion details for the location or size of the lesion To generate diagnostic information including one or more of the information, integrated medical diagnosis service providing method. In the integrated medical diagnosis service providing system that provides diagnosis information through medical image analysis,
A communication unit that processes transmission and reception of signals related to medical images through a communication network;
A memory in which a program for performing an integrated medical diagnosis service providing method based on the medical image analysis is recorded; And
And a control unit for executing the program,
The control unit, by executing the program,
Collect and store unstructured raw medical image data, provide standardized medical image data through pre-processing of the raw medical image data, and extract features of an anatomical structure or lesion of the body from the medical image data. Learning to provide a classification model, extracting features of newly input medical image data based on the classification model,
Deriving a lesion for a corresponding medical image using the extracted features to provide diagnostic information, and inferring at least one suspicious lesion area that is a major factor in deriving the lesion in the medical image based on the diagnostic information Afterwards, visual information of the lesion area is provided using the inferred lesion suspicious area,
After generating at least one keyword matching the anatomy of the body for the lesion region using the visual information of the lesion region, a diagnostic report is generated and provided using the diagnostic information, visual information of the lesion region, and keywords. ,
The visual information includes images obtained by dividing the medical image data, and the keyword is a name of an anatomical structure of the body that matches an image including the lesion suspicious region among the images divided by the medical image data , Integrated medical diagnosis service delivery system. The method of claim 11,
The control unit outputs a diagnosis report including visual information of the lesion area through a display unit. The method of claim 11,
The control unit is to learn a weight for classification of the medical image by using a deep learning algorithm based on CNN (Convolutional Neural Networks), integrated medical diagnosis service providing system. The method of claim 13,
The control unit learns the classification model using a 3D CNN to extract features of medical image data including 3D spatial information. The method of claim 13,
The control unit,
Based on the deep learning algorithm, the medical image data is divided into a preset number of subject images according to the anatomy of the body, and then a label is assigned using a human anatomy atlas for each subject image. Do the labeling work,
Analyzing the characteristics of the anatomical structure of the divided subject image, detecting at least one subject image having a suspicious lesion area, and then deriving a major lesion area using weights,
The system for providing an integrated medical diagnosis service to extract an anatomical name for a body part as a keyword by matching a label assigned to the subject image corresponding to the derived major lesion area. The method of claim 11,
The control unit generates text information for explaining a medical diagnosis by integrating the extracted keyword and diagnosis information, and includes it in the diagnosis report. The method of claim 11,
The control unit,
By inferring at least one suspicious area of the lesion that is a major factor in deriving the lesion through backtracking of the process of deriving the diagnostic information, and projecting the inferred suspicious area of the lesion onto an image plane of the same dimension as the corresponding medical image To generate visual information of the lesion area, integrated medical diagnosis service providing system. The method of claim 11,
The diagnostic information,
Including at least one of lesion presence information indicating the presence or absence of a lesion in the medical image, diagnosis name information through pathology diagnosis of the lesion, diagnosis probability information including a matching probability value for the corresponding diagnosis name, and lesion detail information regarding the location or size of the lesion. That is, an integrated medical diagnosis service providing system.

Images