Method, program, and device for diagnosing myocardial infarction using electrocardiogram

Abstract

According to one embodiment of the present disclosure, disclosed are a method, program, and apparatus for diagnosing myocardial infarction using an electrocardiogram performed by a computing device. The method may comprise: a step of generating learning data from electrocardiogram data of a patient subjected to coronary arteriography; and a step of learning a deep learning model for predicting the onset or progression process of myocardial infarction in the patient using the generated learning data.

Description

Technical Field

[0001] This disclosure relates to artificial intelligence technology in the medical field, specifically to a method of learning a model for diagnosing myocardial infarction using an electrocardiogram and utilizing the learned model. Background Technology

[0002] An electrocardiogram (ECG) is an examination that records the electrical activity of the heart. ECG is a relatively simple and cost-effective method for assessing the health of the heart, playing a crucial role in the early diagnosis and management of heart disease. For example, the ECG signals measured can be used to confirm whether different parts of the heart are functioning normally, whether the heart's size and position are normal, and whether the heart muscle is damaged. Furthermore, based on these findings, ECG signals can be used to diagnose various heart-related problems and predict a person's health status.

[0003] Furthermore, with the development of artificial intelligence (AI) technology, there are increasing attempts to analyze electrocardiograms (ECGs) using AI. In particular, there are many studies on models for diagnosing arrhythmias using ECGs. However, unlike arrhythmias, there are almost no studies on models for diagnosing myocardial infarction. Moreover, it is reported that myocardial infarctions, with approximately 80% to 90% coronary artery blockage, are difficult for even medical professionals to detect using only ECGs. Therefore, AI technology is needed in the field of diagnosing myocardial infarction using ECGs. Summary of the Invention

[0004] Technical issues

[0005] The purpose of this disclosure is to provide a method for learning a model for diagnosing myocardial infarction using electrocardiograms and for diagnosing myocardial infarction using the learned model.

[0006] However, the problems to be addressed in this disclosure are not limited to those mentioned above, and other unmentioned problems can be clearly understood from the following description.

[0007] Technical solution

[0008] According to one embodiment of this disclosure for achieving the objectives described above, a method for diagnosing myocardial infarction using an electrocardiogram (ECG) is disclosed, executed by a computing device. The method may include: generating learning data based on ECG data from a patient undergoing coronary angiography; and learning a deep learning model using the generated learning data to predict the onset or progression of myocardial infarction in the patient.

[0009] Optionally, the step of generating learning data based on the electrocardiogram data of a patient undergoing coronary angiography may include: marking the degree of coronary artery stenosis on the electrocardiogram data of the patient undergoing the coronary angiography to generate learning data.

[0010] Optionally, the marking can be performed based on the type of coronary artery.

[0011] Optionally, the step of learning a deep learning model for predicting the onset or progression of myocardial infarction using the generated learning data may include: classifying the learning data to generate subgroups based on at least one of a first criterion representing demographics, a second criterion representing factors influencing the onset or progression of myocardial infarction, a third criterion representing the type of disease and the subject of the disease, or a fourth criterion representing the type of electrocardiogram; and learning the deep learning model in a manner that predicts the onset or progression of myocardial infarction according to the generated subgroups.

[0012] Optionally, the step of learning a deep learning model for predicting the onset or progression of myocardial infarction using the generated learning data may include: filtering groups with performance below a standard value based on validation results derived during the learning process of the generated subgroups; and performing further learning on the deep learning model that performed the learning based on the filtered groups.

[0013] Optionally, the further learning can be performed based on few-shot learning.

[0014] Optionally, the second group may include at least one of the following: risk factors for the onset of the disease, past medical history, whether the patient is an early-onset patient, or type of chest pain.

[0015] Optionally, the third group may include at least one of the following: types of myocardial infarction, types of acute coronary syndrome, or types of coronary arteries.

[0016] According to one embodiment of this disclosure for achieving the aforementioned objectives, a method for diagnosing myocardial infarction using an electrocardiogram (ECG) is disclosed, executed by a computing device. The method may include: the step of acquiring ECG data; and the step of inputting the ECG data into a pre-learned deep learning model to predict the onset or progression of myocardial infarction in a subject detecting the ECG data. In this case, the deep learning model is pre-learned in a manner that predicts the onset or progression of myocardial infarction in a patient using learning data generated from ECG data of a patient undergoing coronary angiography.

[0017] According to one embodiment of this disclosure for achieving the aforementioned objectives, a computer program stored in a computer-readable storage medium is disclosed. When the computer program is executed on one or more processors, it performs actions to diagnose myocardial infarction using an electrocardiogram (ECG). These actions may include: generating learning data based on ECG data from a patient undergoing coronary angiography; and learning a deep learning model using the generated learning data to predict the onset or progression of myocardial infarction in the patient.

[0018] According to one embodiment of this disclosure for achieving the aforementioned objectives, a computational device for diagnosing myocardial infarction using electrocardiogram (ECG) is disclosed. The device may include: a processor including at least one core; memory including program code executable on the processor; and a network unit for acquiring ECG data. In this case, the processor may generate learning data based on ECG data from a patient undergoing coronary angiography, and use the generated learning data to learn a deep learning model for predicting the onset or progression of myocardial infarction in the patient.

[0019] Invention Effects

[0020] According to the method disclosed herein, an artificial intelligence model can be used to analyze not only whether a person has myocardial infarction, but also the progression of myocardial infarction.

[0021] Furthermore, the shortcomings of models that perform poorly in specific patient populations can be compensated for by learning from the AI ​​model with few shots. Attached Figure Description

[0022] Figure 1 This is a block diagram of a computing device according to an embodiment of the present disclosure.

[0023] Figure 2 A block diagram illustrating the process of learning a deep learning model according to an embodiment of the present disclosure.

[0024] Figure 3 A block diagram illustrating the process of learning a deep learning model according to an additional embodiment of the present disclosure.

[0025] Figure 4 This is a flowchart illustrating a learning method of a deep learning model according to an embodiment of the present disclosure.

[0026] Figure 5 A flowchart illustrating the inference method of a deep learning model according to an embodiment of the present disclosure. Detailed Implementation

[0027] The embodiments of this disclosure are described in detail below with reference to the accompanying drawings, so that those skilled in the art (hereinafter referred to as "those skilled in the art") can readily implement this disclosure. The embodiments presented in this disclosure are intended for use or implementation by those skilled in the art. Therefore, various modifications to the embodiments of this disclosure will be apparent to those skilled in the art. That is, this disclosure can be implemented in many different forms and is not limited to the following embodiments.

[0028] Throughout this disclosure, identical or similar reference numerals refer to identical or similar components. Furthermore, for clarity, reference numerals for parts unrelated to the description of this disclosure may be omitted from the accompanying drawings.

[0029] The term "or" as used in this disclosure is an inclusive "or," not an exclusive "or." That is, unless otherwise specified in this disclosure or its meaning is unclear in the context, "X utilizes A or B" should be understood as referring to one of the natural inclusive substitutions. For example, unless otherwise specified in this disclosure or its meaning is unclear in the context, "X utilizes A or B" could be interpreted as X utilizes A, or X utilizes B, or X utilizes both A and B.

[0030] The term “and / or” as used in this disclosure should be understood to refer to and include all possible combinations of more than one of the related concepts listed.

[0031] The terms “comprising” and / or “including…” as used in this disclosure should be understood to mean the presence of a specific feature and / or component. However, the terms “comprising” and / or “including…” should be understood not to exclude the presence or addition of more than one other feature, other component, and / or combination thereof.

[0032] Unless otherwise specified in this disclosure or explicitly referred to in the context, the singular is generally interpreted as including "one or more".

[0033] The term "Nth (where N is a natural number)" as used in this disclosure is understood to refer to components of this disclosure as distinguished according to criteria such as functional viewpoint, structural viewpoint, or ease of description. For example, components that perform different functions in this disclosure may be distinguished as a first component or a second component. However, components that are substantially the same within the technical concept of this disclosure but need to be distinguished for ease of description may also be distinguished as a first component or a second component.

[0034] As used in this disclosure, the term "acquisition" refers not only to receiving data via a wired or wireless communication network with an external device or system, but also to generating data in an on-device manner.

[0035] Furthermore, the terms "module" or "unit" as used in this disclosure can be understood as referring to an independent functional unit that processes computing resources, such as a computer-related entity, firmware, software or a part thereof, hardware or a part thereof, or a combination of software and hardware. In this case, "module" or "unit" can be a unit composed of a single element, or a unit represented by a combination or set of multiple elements. For example, in a narrow sense, "module" or "unit" can refer to a hardware element of a computing device or a set thereof, an application program that performs a specific function of software, a procedure implemented by running software, or a set of commands for running a program, etc. In a broad sense, "module" or "unit" can refer to the computing device itself that constitutes a system, or an application program running on a computing device, etc. However, the concepts described are merely illustrative, and the concept of "module" or "unit" can be defined in various ways within the scope understandable to those skilled in the art, based on the content of this disclosure.

[0036] As used in this disclosure, the term "model" can be understood as a system implemented using mathematical concepts and language to solve a specific problem, a collection of software units used to solve a specific problem, or an abstract model of a process for solving a specific problem. For example, a neural network "model" can refer to an entire system implemented by a neural network that learns to have problem-solving capabilities. In this case, the neural network acquires problem-solving capabilities by learning to optimize the parameters of the connecting nodes or neurons. A neural network "model" can include a single neural network or a collection of neural networks composed of multiple neural networks.

[0037] The foregoing descriptions of terminology are intended to aid in understanding this disclosure. Therefore, it should be noted that unless the foregoing terms are explicitly stated as being used to define the content of this disclosure, the content of this disclosure is not intended to limit the technical concept.

[0038] Figure 1 This is a block diagram of a computing device according to an embodiment of the present disclosure.

[0039] The computing device 100 of one embodiment of this disclosure may be a hardware device or part of a hardware device that performs comprehensive data processing and computation, or it may be a software-based computing environment connected by a communication network. For example, the computing device 100 may be a main server that performs intensive data processing functions and shares resources, or it may be a client that shares resources through interaction with the server. In addition, the computing device 100 may also be a cloud system in which multiple servers and clients interact to comprehensively process data. The description described is only one example related to the type of computing device 100, and therefore, the type of computing device 100 may be configured in various ways within the scope that can be understood by those skilled in the art based on the content of this disclosure.

[0040] Reference Figure 1 A computing device 100 according to one embodiment of this disclosure may include a processor 110, a memory 120, and a network unit 130. However, Figure 1 This is merely an example; therefore, computing device 100 may include other configurations for implementing a computing environment. Furthermore, computing device 100 may also include only a portion of the disclosed configurations.

[0041] The processor 110 of one embodiment of this disclosure can be understood as a configuration unit including hardware and / or software for performing computational operations. For example, the processor 110 can perform data processing for machine learning by reading a computer program. The processor 110 can handle computational processes such as processing input data for machine learning, feature extraction for machine learning, and error calculation based on backpropagation. The processor 110 for performing the data processing described above may include a central processing unit (CPU), a general-purpose graphics processing unit (GPGPU), a tensor processing unit (TPU), an application-specific integrated circuit (ASIC), or a field-programmable gate array (FPGA), etc. The type of processor 110 is only an example, and the processor 110 can be configured in various ways within the scope of understanding by those skilled in the art based on the content of this disclosure.

[0042] The processor 110 can generate learning data for a deep learning model used to diagnose myocardial infarction based on the electrocardiogram (ECG) data of patients undergoing coronary angiography. The processor 110 can generate learning data by labeling the degree of coronary artery stenosis on the ECG data of patients undergoing coronary angiography. The degree of coronary artery stenosis, confirmed through the coronary angiography procedure, can be acquired along with the ECG data. In other words, the computing device 100 can generate learning data by inputting the ECG data of patients undergoing coronary angiography and labeling the degree of coronary artery stenosis for the corresponding patients.

[0043] Processor 110 can use the generated learning data to learn a deep learning model for diagnosing myocardial infarction. Processor 110 is capable of learning the deep learning model in a way that allows it to predict the onset or progression of myocardial infarction in a patient based on the learning data. For example, processor 110 can input electrocardiogram data included in the learning data into the deep learning model to output the degree of coronary artery stenosis for predicting the onset or progression of myocardial infarction in a patient. The output is a numerical value representing the degree of coronary artery stenosis, which can be in fractional form. Processor 110 can adjust the parameters of the neural network included in the deep learning model by comparing the output and labels of the deep learning model. Processor 110 can repeatedly execute this process to learn the deep learning model until the error between the output and labels of the deep learning model meets a minimum criterion.

[0044] Processor 110 can predict the onset or progression of myocardial infarction using a learned deep learning model. Processor 110 can input electrocardiogram (ECG) data into the learned deep learning model to output the degree of coronary artery stenosis for predicting the onset or progression of myocardial infarction. The ECG data input to the deep learning model is not limited to data from patients undergoing coronary angiography; it can include all ECG data collected for various purposes in medical settings or in real life. Based on the degree of coronary artery stenosis output by the deep learning model, processor 110 can predict whether myocardial infarction will occur, or if so, to what extent it will progress.

[0045] The memory 120 of one embodiment of this disclosure can be understood as including hardware and / or software configuration units that store and manage data processed in the computing device 100. That is, the memory 120 can store data of any form generated or determined by the processor 110 and data of any form received by the network unit 130. For example, the memory 120 may include at least one type of storage medium selected from flash memory type, hard disk type, multimedia card micro type, card type memory, random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, and optical disk. In addition, the memory 120 may also include a database system that controls and manages data under a defined system. The type of memory 120 described is merely an example; therefore, the type of memory 120 can be configured in various ways within the scope that can be understood by a person skilled in the art based on the content of this disclosure.

[0046] The memory 120 can be managed by structuring and organizing the data required by the processor 110 to perform operations, combinations of data, and executable program code in the processor 110. For example, the memory 120 may store medical data received by the network unit 130, which will be described later. The memory 120 may store program code that enables the neural network model to receive medical data to perform learning, program code that enables the neural network model to receive medical data to perform inference according to the purpose of use of the computing device 100, and processing data generated as the program code is executed.

[0047] Network unit 130 of one embodiment of this disclosure can be understood as a configuration unit that transmits or receives data through any form of known wired or wireless communication system. For example, network unit 130 can transmit or receive data through a wired or wireless communication system using a local area network (LAN), wideband code division multiple access (WCDMA), long term evolution (LTE), wireless broadband internet (WiBro), 5G, ultra-wideband wireless communication, ZigBee, radio frequency communication, wireless LAN, wireless fidelity, near field communication (NFC), or Bluetooth. The communication system described is merely an example; therefore, wired or wireless communication systems used by network unit 130 to transmit or receive data can be applied in various ways besides the example described.

[0048] Network unit 130 can receive data required for computation by processor 110 via wired or wireless communication with any system or client. Furthermore, network unit 130 can transmit data generated by the computation of processor 110 via wired or wireless communication with any system or client. For example, network unit 130 can receive medical data via communication with databases within a hospital environment, cloud servers performing tasks such as standardizing medical data, clients such as smartwatches, or medical computing devices. Network unit 130 can also transmit output data of neural network models and intermediate data and processed data derived during the computation of processor 110 via communication with the aforementioned databases, servers, clients, or computing devices.

[0049] Figure 2 A block diagram illustrating the process of learning a deep learning model according to an embodiment of the present disclosure. Figure 3 A block diagram illustrating the process of learning a deep learning model according to an additional embodiment of the present disclosure.

[0050] Reference Figure 2In one embodiment of this disclosure, the computing device 100 generates learning data 20 by labeling the degree of coronary artery stenosis in the electrocardiogram (ECG) data 10 of a patient undergoing coronary angiography. At this time, the computing device 100 can label the degree of coronary artery stenosis according to the type of coronary artery. For example, the computing device 100 can label the degree of stenosis in the ECG data 10 according to the medical classification of coronary arteries such as the right coronary artery (RCA), left anterior descending artery (LAD), and left circumflex artery (LCx). Through this labeling, the deep learning model can analyze which blood vessels are stenotic to what degree during the learning process.

[0051] The computing device 100 can learn the deep learning model 200 by inputting the learning data 20 into it, thereby outputting a value indicating the degree of coronary artery stenosis for diagnosing myocardial infarction. The deep learning model 200 may include a neural network that extracts features related to the electrocardiogram waveform, such as the P wave, QRS complex, and T wave, and a neural network that outputs a numerical value representing the degree of coronary artery stenosis based on the extracted features. For example, the computing device 100 can output a numerical value representing the degree of coronary artery stenosis by inputting the electrocardiogram data included in the learning data 20 into the deep learning model 200. The computing device 100 can calculate the error using a loss function with the numerical value output by the deep learning model 200 and the labels included in the learning data 20 as input variables. The computing device 100 can adjust the parameters of the neural network included in the deep learning model 200 based on the calculated error. Furthermore, the computing device 100 can repeatedly perform the process of adjusting the neural network parameters until the error meets a minimum criterion. As illustrated, the learning of the deep learning model 200 can be based on guided learning, but it can also be based on self-guided learning, depending on the neural network structure of the deep learning model 200.

[0052] Furthermore, if the deep learning model 200 has not learned many electrocardiograms (ECGs) from patients with myocardial infarction, the accuracy of the output will be low due to insufficient learning of ECG data from these patients. Typically, the number of ECG data from patients with diseases is less than that from healthy individuals. Therefore, there is a possibility that the deep learning model 200 may not have adequately learned the amount of data needed to diagnose diseases such as myocardial infarction. Therefore, to improve the performance of the deep learning model 200, the computing device 100 may perform further learning on data sets with low output accuracy based on validation evaluations of specific data sets identified during the learning process.

[0053] Reference Figure 3 The computing device 100 can classify learning data into multiple subgroups based on factors used in the medical diagnosis of myocardial infarction. For example, the computing device 100 can classify learning data 20 into a first subgroup 31, a second subgroup 33, a third subgroup 35, and an Nth subgroup 39 that respectively match multiple criteria that influence the diagnosis of myocardial infarction. Specifically, the classification criteria corresponding to subgroups 31, 33, 35, and 39 may include a first criterion representing demographics, a second criterion representing factors influencing the onset or progression of myocardial infarction, a third criterion representing the type of disease and the affected individuals, or a fourth criterion representing the type of electrocardiogram. The second criterion may include at least one of the following: risk factors for disease onset, medical history, whether the patient is in the early stage of the disease, or the type of chest pain. Furthermore, the third criterion may include at least one of the following: the type of myocardial infarction, the type of acute coronary syndrome, or the type of coronary artery disease. The computing device 100 can generate subgroups 31, 33, 35, and 39 by classifying the data included in the learning data 20 into one of the first to fourth criteria.

[0054] Table 1

[0055]

[0056]

[0057] The specific details of each standard are summarized in Table 1. However, Table 1 is only an example, and the standards disclosed herein are not limited to those in Table 1.

[0058] The computing device 100 can perform a first learning process, which enables the deep learning model 200 to predict the onset or progression of myocardial infarction according to subgroups. The computing device 100 can perform the first learning process by inputting subgroups into the deep learning model 200 and outputting the degree of coronary artery stenosis for predicting the onset or progression of myocardial infarction. At this time, the first learning process can be combined with the... Figure 2 The learning process corresponds to this.

[0059] The computing device 100 can derive verification results for each subgroup during the execution of the first learning process. Furthermore, the computing device 100 can filter out subgroups with low performance based on the verification results. For example, the computing device 100 can filter out subgroups with performance below a standard value based on the verification result values ​​implemented during the execution of the first learning process. The computing device 100 can also filter out subgroups with performance below a standard value based on performance indicators used in hospitals or emergency rooms.

[0060] The computing device 100 can perform a second learning process on the deep learning model 200 that has completed the first learning, based on the selected groups. The computing device 100 inputs groups whose performance is below a standard value into the deep learning model 200 that has completed the first learning, and performs further learning based on few-shot learning to improve the performance of these groups. Here, few-shot learning is a learning method that performs inference based on a small amount of data; it can be understood as learning the similarities between a small amount of data to learn which classes the query data belongs to in the supporting data. That is, groups whose performance is below the standard value are considered a small amount of data. Therefore, the computing device 100 can perform a second learning process based on few-shot learning, inputting the groups whose performance is below the standard value as a small amount of data into the deep learning model 200 that has completed the first learning to predict the degree of coronary artery stenosis. The computing device 100 can use this second learning to compensate for the performance deficiencies of the deep learning model 200 that show insufficient performance under specific conditions or diagnostic factors.

[0061] Figure 4 This is a flowchart illustrating a learning method of a deep learning model according to an embodiment of the present disclosure.

[0062] Reference Figure 4 In one embodiment of this disclosure, the computing device 100 can generate learning data based on the electrocardiogram (ECG) data of a patient undergoing coronary angiography (step S110). When the computing device 100 is a server for interpreting ECGs, it can acquire the ECG data of the patient undergoing coronary angiography by communicating with a database in a medical environment that manages the patient's ECG data. When the computing device 100 is a server that manages ECG data in a medical environment, if the detection of the ECG signal for the patient undergoing coronary angiography is completed, the computing device 100 can generate and manage the corresponding patient's ECG data. The computing device 100 can generate learning data by marking the degree of stenosis of the coronary arteries of the patient in the ECG data of the patient undergoing coronary angiography. Furthermore, when marking the degree of stenosis of the coronary arteries, the computing device 100 can distinguish and mark the types of coronary arteries.

[0063] The computing device 100 can use the learning data generated in step S110 to learn a deep learning model for predicting the onset or progression of myocardial infarction in patients (step S120). The computing device 100 can learn the deep learning model through a first learning process based on guided learning, self-guided learning, etc., according to the structure of the deep learning model. Furthermore, the computing device 100 can further learn the deep learning model through a second learning process based on few-shot learning, identifying datasets with low prediction accuracy of the deep learning model. For this learning, the computing device 100 can classify and use the learning data in multiple ways according to factors related to myocardial infarction diagnosis. Specifically, the computing device 100 can perform the first learning process by inputting subgroups generated from the classification of the learning data into the deep learning model. The computing device 100 can filter out groups with low learning performance based on the validation results of each subgroup confirmed through the first learning process. Furthermore, the computing device 100 can perform the second learning process on the deep learning model based on the filtered groups.

[0064] Figure 5 A flowchart illustrating the inference method of a deep learning model according to an embodiment of the present disclosure.

[0065] Reference Figure 5 In one embodiment of this disclosure, the computing device 100 can acquire electrocardiogram (ECG) data (step S210). When the computing device 100 is a server for interpreting ECGs, it can acquire the ECG data of the person whose ECG is to be interpreted by communicating with devices in a medical environment that detect patient ECG data or wearable devices that can detect ECG data. When the computing device 100 is a device in a medical environment that detects patient ECG data or a wearable device that can detect ECG data, it can generate ECG data by detecting ECG signals.

[0066] The computing device 100 can predict the onset or progression of myocardial infarction in an object whose electrocardiogram (ECG) data is being analyzed by inputting ECG data into a pre-learned deep learning model (step S220). The computing device 100 can also output values ​​representing the degree of coronary artery stenosis by inputting ECG data into the pre-learned deep learning model. Based on clinical evidence analysis of the degree of coronary artery stenosis output by the deep learning model, the computing device 100 derives the onset or progression of myocardial infarction. The computing device 100 can generate a report including information about the derived onset or progression of myocardial infarction. In this case, if the computing device 100 is a server for interpreting ECGs, it can share the report by communicating with a client capable of confirming the ECG interpretation results. If the computing device 100 is equipped with an input / output unit providing a user interface, it can output the report through the input / output unit.

[0067] The various embodiments of this disclosure described above can be combined with additional embodiments and can be modified within the scope understandable to those skilled in the art based on the detailed description. It should be understood that the embodiments of this disclosure are illustrative in all respects and not limiting. For example, each component described as a single type can be implemented separately, and similarly, components described as distributed can be implemented in combination. Therefore, it should be explained that the meaning, scope, and all modifications or variations derived from the invention claims of this disclosure are included within the scope of this disclosure.

Claims

1. A method for diagnosing myocardial infarction using an electrocardiogram, performed by a computing device comprising at least one processor, characterized in that: include: The step of generating learning data based on electrocardiogram data of a patient undergoing coronary angiography; as well as The step of using the generated learning data to learn a deep learning model for predicting the onset or progression of myocardial infarction in the patient.

2. The method according to claim 1, characterized in that Based on the electrocardiogram data of patients undergoing coronary angiography, the steps for generating learning data include: The step of generating learning data by marking the degree of coronary artery stenosis on electrocardiogram data of a patient who has undergone the coronary angiography.

3. The method according to claim 2, characterized in that The marking is performed by distinguishing the coronary arteries according to their types.

4. The method according to claim 3, characterized in that The steps of learning a deep learning model for predicting the onset or progression of myocardial infarction using the generated learning data include: A step of classifying the learning data to generate subgroups based on at least one of a first criterion representing demographics, a second criterion representing factors that affect the onset or progression of myocardial infarction, a third criterion representing the types of diseases and targets, or a fourth criterion representing electrocardiogram types; and The step of learning the deep learning model in a manner that predicts the onset or progression of the myocardial infarction according to the generated subset.

5. The method according to claim 3, characterized in that: The steps of learning a deep learning model for predicting the onset or progression of myocardial infarction using the generated learning data include: A step of selecting a group having performance below a standard value based on a verification result derived during a learning process of the generated subgroup; and Based on the filtered group, a further learning step is performed on the deep learning model that performs the learning.

6. The method according to claim 5, characterized in that The further learning is performed based on few-shot learning.

7. The method according to claim 1, characterized in that The second criterion includes at least one of the risk factors for the onset of the disease, previous history, whether the patient is an early-stage patient, or the type of chest pain.

8. The method according to claim 1, characterized in that The third criterion includes at least one of a type of myocardial infarction, a type of acute coronary syndrome, or a type of coronary artery.

9. A method for diagnosing myocardial infarction using an electrocardiogram, performed by a computing device including at least one processor, characterized in that: include: Steps for acquiring electrocardiogram data; and The electrocardiogram data is input into a pre-learned deep learning model to predict the onset or progression of myocardial infarction in a subject detecting the electrocardiogram data, The deep learning model is pre-learned in a manner that predicts the onset or progression of myocardial infarction in a patient who undergoes coronary angiography using learning data generated based on electrocardiogram data of the patient.

10. A computer program stored in a computer-readable storage medium, characterized in that: When the computer program is executed by one or more processors, the computer program performs the steps of diagnosing myocardial infarction using electrocardiogram. The actions include: An action of generating learning data based on electrocardiogram data of a patient undergoing coronary angiography; and Using the generated learning data, the operation of a deep learning model for predicting the onset or progression of myocardial infarction in the patient is learned.

11. A computing device for diagnosing myocardial infarction using an electrocardiogram, characterized in that: include: a processor, comprising at least one core; a memory including program code executable on the processor; and A network unit for acquiring electrocardiogram data, The processor generates learning data based on electrocardiogram data of a patient undergoing coronary angiography, and uses the generated learning data to learn a deep learning model for predicting the onset or progression of myocardial infarction in the patient.

Images