WO2024101759A1

WO2024101759A1 - Method, program, and device for analyzing electrocardiogram data

Info

Publication number: WO2024101759A1
Application number: PCT/KR2023/017193
Authority: WO
Inventors: 권준명
Original assignee: 주식회사 메디컬에이아이
Priority date: 2022-11-09
Filing date: 2023-11-01
Publication date: 2024-05-16
Also published as: KR20240067480A

Abstract

A method for analyzing electrocardiogram data, according to one embodiment of the present disclosure, may comprise the steps of: acquiring an electrocardiogram signal; generating electrocardiogram feature data including a plurality of features on the basis of the electrocardiogram signal; and generating main electrocardiogram component data by reducing the dimension of the electrocardiogram feature data on the basis of one or more main components. The one or more main components are characterized in that the plurality of features is merged on the basis of the correlation between the plurality of features.

Description

Methods, programs and devices for analysis of electrocardiogram data

The present disclosure relates to data processing technology in the medical field, and specifically relates to a method of facilitating analysis by reducing the dimension of feature data extracted from electrocardiogram signals.

An electrocardiogram is a graphical recording of potentials related to heartbeat on the surface of the body. In addition to the standard 12-lead electrocardiogram, there is an exercise stress electrocardiogram and an electrocardiogram during activity. Electrocardiograms are used for screening and diagnosis of circulatory diseases and can be obtained easily. In addition, the electrocardiogram has the advantage of being relatively inexpensive, non-invasive, and easily repeatable recording.

Referring to FIG. 1, a plurality of features can be extracted from the ECG signal through analysis. An electrocardiogram refers to a recording of electrical signals occurring in the heart, and can usually be measured at a length of about 10 seconds. The heart is recorded beating several times in 10 seconds, and the recording of one beat of the heart can be called 1 beat. One beat of an electrocardiogram may include a P wave corresponding to atrial depolarization, a QRS wave corresponding to ventricular depolarization, and a T wave corresponding to ventricular repolarization. The extracted features are 'PR INTERVAL', which is the interval from the start point of the P wave to the start point of the QRS wave, 'QRS duration', which is the interval of the QRS wave, 'P AREA', which is the area of the + area among the P waves, and the + value among the P waves. This may include 'P AMP', which is the voltage value of the highest point.

Users can analyze the patient's health status based on features extracted from the electrocardiogram signal. However, when the number of features is one, it is easy to understand its meaning, but when the number of features exceeds dozens, it is not easy to synthesize the features and come to a conclusion. Moreover, if there are hundreds or tens of millions of ECG data, not just one person's, and you try to synthesize and analyze them, as well as their relationships with other data, it becomes even more difficult.

Therefore, there is a need for a method for converting ECG signals or characteristic data of ECG signals into a form for effective analysis.

The present disclosure was developed in response to the above-described background technology, and its purpose is to provide a method of reducing the dimension of ECG feature data extracted from ECG signals and converting them into data in a form that is easy to analyze.

However, the problems to be solved by this disclosure are not limited to the problems mentioned above, and other problems not mentioned can be clearly understood based on the description below.

A method of analyzing electrocardiogram data, which is performed by a computing device, is disclosed according to an embodiment of the present disclosure for realizing the above-described problem. The method includes obtaining an electrocardiogram signal, generating electrocardiogram feature data including a plurality of features based on the electrocardiogram signal, and reducing the dimension of the electrocardiogram feature data based on one or more principal components to obtain electrocardiogram principal component data. It may include the step of generating. The one or more main components are characterized in that the plurality of features are merged based on the correlation between the plurality of features.

Alternatively, generating the ECG feature data may include extracting n features (n is a natural number) based on a preset feature extraction algorithm.

Alternatively, generating the ECG feature data may include inferring n features by inputting the ECG signal into a pre-trained first neural network model.

Alternatively, the principal component may be characterized as including a first principal component corresponding to the dimensional axis in which the variance of the ECG principal component data is greatest.

Alternatively, the method may further include visualizing the first principal component and the ECG principal component data in a first graph.

Alternatively, the first graph may include first disease area information.

Alternatively, the main component may further include a second main component orthogonal to the first main component.

Alternatively, the method may further include visualizing the ECG feature data in a second graph based on the first principal component and the second principal component.

Alternatively, the step of predicting a disease corresponding to the ECG main component data by inputting the ECG main component data into a pre-trained second neural network model may be further included.

According to an embodiment of the present disclosure for realizing the above-described object, a computer program stored in a computer-readable storage medium is disclosed. When the computer program runs on one or more processors, it performs operations for analyzing electrocardiogram data. At this time, the operations include acquiring an ECG signal, generating ECG feature data including a plurality of features based on the ECG signal, and reducing the dimension of the ECG feature data based on one or more main components, An operation for generating principal component data may be included. The one or more main components are characterized in that the plurality of features are merged based on the correlation between the plurality of features.

According to an embodiment of the present disclosure for realizing the above-described problem, a computing device for inputting electrocardiogram data through an image projection method is disclosed. The device includes a processor including at least one core; a memory containing program codes executable on the processor; And it may include a network unit for acquiring electrocardiogram data. At this time, the processor acquires an ECG signal, generates ECG feature data including a plurality of features based on the ECG signal, reduces the dimension of the ECG feature data based on one or more main components, and generates ECG main component data. creates . The one or more main components are characterized in that the plurality of features are merged based on the correlation between the plurality of features.

The present disclosure can convert ECG feature data including a plurality of features into a form that is easy to analyze by using a data dimension reduction method including principal component analysis.

The present disclosure enables intuitive analysis by visualizing ECG feature data with reduced dimensionality.

The present disclosure can predict a user's disease based on dimensionally reduced ECG feature data.

1 is a diagram explaining ECG feature data extracted from an ECG signal.

2 is a block diagram of a computing device according to an embodiment of the present disclosure.

Figure 3 is a block diagram illustrating an electrocardiogram data analysis device according to an embodiment of the present disclosure.

Figure 4 is a block diagram illustrating a neural network model according to an embodiment of the present disclosure.

FIG. 5 is a diagram illustrating a method of reducing the dimension of ECG feature data in the ECG data analysis method according to an embodiment of the present disclosure.

Figure 6 is a diagram visualizing the result of reducing the dimension of ECG data according to an embodiment of the present disclosure.

FIG. 7 is a diagram illustrating a method for visualizing ECG data according to an embodiment of the present disclosure.

Figure 8 is a block diagram illustrating a neural network model according to an embodiment of the present disclosure.

9 is a flowchart explaining an ECG data analysis method according to an embodiment of the present disclosure.

Figure 10 is a flowchart explaining an ECG data analysis method according to an embodiment of the present disclosure.

Below, with reference to the attached drawings, embodiments of the present disclosure are described in detail so that those skilled in the art (hereinafter referred to as skilled in the art) can easily practice the present disclosure. The embodiments presented in this disclosure are provided to enable any person skilled in the art to use or practice the subject matter of this disclosure. Accordingly, various modifications to the embodiments of the present disclosure will be apparent to those skilled in the art. That is, the present disclosure can be implemented in various different forms and is not limited to the following embodiments.

The same or similar reference numerals refer to the same or similar elements throughout the specification of this disclosure. Additionally, in order to clearly describe the present disclosure, reference numerals of parts in the drawings that are not related to the description of the present disclosure may be omitted.

As used in this disclosure, the term “or” is intended to mean an inclusive “or” and not an exclusive “or.” That is, unless otherwise specified in the present disclosure or the meaning is not clear from the context, “X uses A or B” should be understood to mean one of natural implicit substitutions. For example, unless otherwise specified in the present disclosure or the meaning is not clear from the context, “X uses A or B” means that It can be interpreted as one of the cases where all B is used.

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

The terms “comprise” and/or “comprising” as used in this disclosure should be understood to mean that certain features and/or elements are present. However, the terms "comprise" and/or "including" should be understood as not excluding the presence or addition of one or more other features, other components, and/or combinations thereof.

Unless otherwise specified in this disclosure or the context is clear to indicate a singular form, the singular should generally be construed to include “one or more.”

The term “Nth (N is a natural number)” used in the present disclosure can be understood as an expression used to distinguish the components of the present disclosure according to a predetermined standard such as a functional perspective, a structural perspective, or explanatory convenience. there is. For example, in the present disclosure, components performing different functional roles may be distinguished as first components or second components. However, components that are substantially the same within the technical spirit of the present disclosure but must be distinguished for convenience of explanation may also be distinguished as first components or second components.

The term “acquisition” used in this disclosure is understood to mean not only receiving data through a wired or wireless communication network with an external device or system, but also generating data in an on-device form. It can be.

Meanwhile, the term "module" or "unit" used in this disclosure refers to a computer-related entity, firmware, software or part thereof, hardware or part thereof. , can be understood as a term referring to an independent functional unit that processes computing resources, such as a combination of software and hardware. At this time, the “module” or “unit” may be a unit composed of a single element, or may be a unit expressed as a combination or set of multiple elements. For example, a "module" or "part" in the narrow sense is a hardware element or set of components of a computing device, an application program that performs a specific function of software, a process implemented through the execution of software, or a program. It can refer to a set of instructions for execution, etc. Additionally, as a broad concept, “module” or “unit” may refer to the computing device itself constituting the system, or an application running on the computing device. However, since the above-described concept is only an example, the concept of “module” or “unit” may be defined in various ways within a range understandable to those skilled in the art based on the contents of the present disclosure.

As used in this disclosure, the term "model" refers to a system implemented using mathematical concepts and language to solve a specific problem, a set of software units to solve a specific problem, or a process to solve a specific problem. It can be understood as an abstract model of a process. For example, a neural network “model” may refer to an overall system implemented as a neural network that has problem-solving capabilities through learning. At this time, the neural network can have problem-solving capabilities by optimizing parameters connecting nodes or neurons through learning. A neural network “model” may include a single neural network or a neural network set in which multiple neural networks are combined.

The explanation of the foregoing terms is intended to aid understanding of the present disclosure. Therefore, if the above-mentioned terms are not explicitly described as limiting the content of the present disclosure, it should be noted that the content of the present disclosure is not used in the sense of limiting the technical idea.

The computing device 100 according to an embodiment of the present disclosure may be a hardware device or part of a hardware device that performs comprehensive processing and calculation of data, or may be a software-based computing environment connected to a communication network. For example, the computing device 100 may be a server that performs intensive data processing functions and shares resources, or it may be a client that shares resources through interaction with the server. Additionally, the computing device 100 may be a cloud system in which a plurality of servers and clients interact to comprehensively process data. Since the above description is only an example related to the type of computing device 100, the type of computing device 100 may be configured in various ways within a range understandable to those skilled in the art based on the contents of the present disclosure.

The computing device 100 according to an embodiment of the present disclosure may include a processor 110, a memory 120, and a network unit 130. However, since FIG. 2 is only an example, the computing device 100 may include other components for implementing a computing environment. Additionally, only some of the configurations disclosed above may be included in computing device 100.

The processor 110 according to an embodiment of the present disclosure may be understood as a structural unit including hardware and/or software for performing computing operations. For example, the processor 110 may read a computer program to extract features of an ECG signal and perform principal component analysis on the extracted features. Additionally, the processor 110 can read a computer program and perform data processing for machine learning. The processor 110 can process computational processes such as processing input data for machine learning, extracting features for machine learning, and calculating errors based on backpropagation. The processor 110 for performing such data processing includes a central processing unit (CPU), a general purpose graphics processing unit (GPGPU), a tensor processing unit (TPU), and a custom processing unit (TPU). It may include a semiconductor (ASIC: application specific integrated circuit), or a field programmable gate array (FPGA: field programmable gate array). Since the type of processor 110 described above is only an example, the type of processor 110 may be configured in various ways within a range understandable to those skilled in the art based on the contents of the present disclosure.

The processor 110 may generate ECG feature data based on the ECG signal. The processor 110 may reduce the dimension of ECG feature data. Additionally, the processor 110 may visualize the dimensionally reduced ECG feature data or predict the health status of the ECG measurement target using the dimensionally reduced ECG feature data.

For example, the processor 110 may learn a deep learning model that extracts ECG features by using ECG signals as input. Additionally, the processor 110 may use the learned deep learning model to generate ECG feature data including a user-specified number of ECG features based on the ECG signal, which is the inference target data. The processor 110 may extract ECG features using a rule-based algorithm.

The processor 110 may reduce the dimensionality of ECG feature data using principal component analysis. Additionally, the processor 110 can visualize the ECG feature data with reduced dimensions to be expressed in a virtual space. Through this data dimension reduction and visualization, the processor 110 can process the data into a form that allows even people without expert knowledge of electrocardiograms to intuitively understand the analysis results. In addition, the processor 110 can learn a deep learning model that diagnoses the type, onset, and course of a disease for the measurement target or predicts the onset of the disease, based on electrocardiogram data whose dimensions have been reduced through principal component analysis. there is. The processor 110 can use dimensionality reduction data to enable a deep learning model to easily perform data interpretation and learning. In addition, the processor 110 can predict, diagnose, or classify the health status of the subject of electrocardiogram measurement based on the dimensionality reduction data that is the subject of inference using the learned deep learning model.

The memory 120 according to an embodiment of the present disclosure may be understood as a structural unit including hardware and/or software for storing and managing data processed in the computing device 100. That is, the memory 120 can store any type of data generated or determined by the processor 110 and any type of data received by the network unit 130. For example, the memory 120 may be a flash memory type, hard disk type, multimedia card micro type, card type memory, or random access memory (RAM). ), SRAM (static random access memory), ROM (read-only memory), EEPROM (electrically erasable programmable read-only memory), PROM (programmable read-only memory), magnetic memory , may include at least one type of storage medium among a magnetic disk and an optical disk. Additionally, the memory 120 may include a database system that controls and manages data in a predetermined system. Since the type of memory 120 described above is only an example, the type of memory 120 may be configured in various ways within a range understandable to those skilled in the art based on the contents of the present disclosure.

The memory 120 can manage data required for the processor 110 to perform operations, a combination of data, and program code executable on the processor 110 by structuring and organizing them. For example, the memory 120 may store ECG data received through the network unit 130, which will be described later. The memory 120 includes program code for mathematical operations that generate ECG feature data and perform dimensionality reduction, program code that operates the neural network model to perform learning by receiving ECG data, and a computing device that receives ECG data from the neural network model. Program code that operates to perform inference according to the purpose of use of (100) and processed data generated as the program code is executed may be stored.

The network unit 130 according to an embodiment of the present disclosure may be understood as a structural unit that transmits and receives data through any type of known wired or wireless communication system. For example, the network unit 130 is a local area network (LAN), wideband code division multiple access (WCDMA), long term evolution (LTE), and WiBro (wireless). broadband internet, 5th generation mobile communication (5G), ultra wide-band wireless communication, ZigBee, radio frequency (RF) communication, wireless LAN, wireless fidelity ), data transmission and reception can be performed using a wired or wireless communication system such as near field communication (NFC), or Bluetooth. Since the above-described communication systems are only examples, the wired and wireless communication systems for data transmission and reception of the network unit 130 may be applied in various ways other than the above-described examples.

The network unit 130 may receive data necessary for the processor 110 to perform operations through wired or wireless communication with any system or client. Additionally, the network unit 130 may transmit data generated through the calculation of the processor 110 through wired or wireless communication with any system or any client. For example, the network unit 130 may receive ECG data through communication with a database in a hospital environment, a cloud server that performs tasks such as standardization of medical data, or the computing device 100. The network unit 130 transmits the output data of the neural network model, intermediate data derived from the calculation process of the processor 110, processed data, etc. through communication with the above-described database, server, or computing device 100. You can.

Referring to FIG. 3 , the ECG data analysis device 200 may include an input module 210, a feature extraction module 220, and a main component data generation module 230. The ECG data analysis device 200 may further include a visualization module 240 or an analysis module 250.

The input module 210 may be an electrocardiogram measurement device that directly measures the user's electrocardiogram signal or a communication module for receiving the electrocardiogram signal transmitted through a network. If the input module 210 is a communication module for receiving an electrocardiogram signal, the input module 210 may correspond to the network unit 130 of FIG. 1.

The feature extraction module 220 may extract one or more features from the user's ECG signal based on a preset feature extraction algorithm. In an additional embodiment, the feature extraction module 220 may include a first neural network model for inferring n features by inputting an ECG signal. For example, the first neural network model may be learned based on the ECG signal and a plurality of features extracted from the ECG signal using a preset algorithm.

The principal component data generation module 230 may reduce the dimension of ECG feature data and generate ECG principal component data based on principal component analysis. For example, the principal component data generation module 230 may convert a plurality of feature variables into one principal component. The specific principal component analysis method is explained in FIG. 5.

The visualization module 240 may display ECG main component data including one, two, or three main components on a graph and provide it to the user. The specific visualization method is explained in Figure 7.

The analysis module 250 can predict diseases corresponding to ECG main component data using statistical methods. In an additional embodiment, the analysis module 250 may include a second neural network model learned with ECG principal component data and disease information determined based on the ECG principal component data. The second neural network model for predicting the user's disease can generate disease prediction data by inputting electrocardiogram main component data.

Referring to FIG. 4, the neural network model 320 may be learned based on the ECG signal and a plurality of features extracted from the ECG signal by a preset algorithm. The neural network model 320 for inferring the characteristics of the ECG signal may generate ECG characteristic data 330 by inputting the measured ECG signal 310.

For example, the neural network model 320 can input electrocardiogram data as learning data into a convolutional neural network, extract and interpret electrocardiogram features from the electrocardiogram data. The neural network model 320 can calculate the loss between the analysis result of the convolutional neural network and the ground truth (GT) using loss functions such as mean square error (MSE), cross entropy, etc. Additionally, the neural network model 320 can optimize the parameters of the convolutional neural network by minimizing loss through optimization techniques such as backpropagation and stochastic gradient descent (SGD). The neural network model 320 can output a numerical value regarding the future health state based on the standardized data that is the subject of inference using the convolutional neural network learned in this way. The type of neural network described above is only an example, and the type of neural network can be configured in various ways, such as a recurrent neural network and a multi-layer perceptron (MLP) neural network, depending on the characteristics of the ECG data.

Referring to FIG. 5, the ECG data analysis method can reduce the dimension of ECG feature data using Principal Component Analysis (PCA). For example, ECG feature data may include a first feature (FT1) and a second feature (FT2). The ECG data analysis method can convert the first feature (FT1) and the second feature (FT2) into the first principal component (PC1), which is a new linearly independent variable.

Graph (a) is a graph showing ECG feature data with the first feature (FT1) on the X-axis and the first feature (FT1) on the Y-axis. The first ECG feature data (DT_1), the second ECG feature data (DT_2), and the Nth ECG feature data (DT_N) are graphed (a) according to the first feature (FT1) and second feature (FT2) values, respectively. ) can be displayed.

The first principal component (PC1) axis displayed in graph (a) may be the axis where the dispersion of the first ECG feature data (DT_1), the second ECG feature data (DT_2), and the Nth ECG feature data (DT_N) is preserved to a maximum. The first principal component (PC1) may be one variable that contains as much of the information contained in the two variables as the first feature (FT1) and the second feature (FT2).

The ECG data analysis method uses principal component analysis to convert first ECG feature data (DT_1) into first ECG principal component data (PDT_1) and convert second ECG feature data (DT_2) into second ECG principal component data (PDT_2). It can be understood in this way.

The converted ECG principal component data can be displayed in graph (b). ECG principal component data may include only one first principal component (PC1) variable, which is a variable that includes all information of the ECG feature data. Therefore, users can intuitively determine disease status, etc. through graph (b).

In FIG. 5, the process of converting two variables into one main component is explained, but the present disclosure is not limited to this and provides M ECG feature data containing N variables (N is an integer greater than 1) (M is It can be converted to the principal components of an integer greater than 0 and less than N. For example, when the ECG feature data includes three or more features, a second main component may be extracted in addition to the first main component. The first principal component may be in the form of an eigenvector having a direction with the greatest variance, and the second principal component may be in the form of an eigenvector that is orthogonal to the first principal component and has the next largest variance after the first principal component.

The ECG data analysis device can enable intuitive and easy interpretation of the ECG by reducing the dimension of the highly complex ECG characteristic data.

Referring to FIG. 6, Table (a) is an example of ECG feature data, and Table (b) may be an example of principal component data obtained by reducing the dimension of the ECG feature data in Table (a) using principal component analysis.

The first feature (FT1) included in table (a) is heart rate, the second feature (FT2) is QRS duration, the third feature (FT3) is PR interval, and the fourth feature (FT4) is ) may be the QRS axis. The first data (DATA1), second data (DATA2), and third data (DATA3) included in table (a) may be ECG characteristic data of different users, and ECG characteristic data measured at different times of the same user. It may be.

The first principal component (PC1) included in Table (b) is a variable that combines the first feature (FT1), the second feature (FT2), the third feature (FT3), and the fourth feature (FT4), and is a variable that combines the first feature (FT1), the second feature (FT2), the third feature (FT3), and the fourth feature (FT4). It may be a variable that maximizes variance to include as much information as possible. The second principal component (PC2) included in Table (b) may be a variable of a component orthogonal to the first principal component (PC1).

Table (b) illustrates two main components by way of example, but in the ECG data analysis method, only one main component may be extracted, or two or more main components may be extracted, considering the features included in the ECG characteristic data.

6 and 7, the ECG data analysis device extracts the first main component (PC1) from the ECG characteristic data and displays first data (DATA1), second data (DATA2), and third data ( DATA3) can be displayed. Graph (a) may be a one-dimensional graph with the first principal component (PC1) as a variable. The ECG data analysis device may set a first region (MI) where the first disease (eg, myocardial infarction) occurs with a high probability through statistical analysis or a preset algorithm. The user may determine that the third data (DATA3) in graph (a) has a high probability of corresponding to the first disease.

The ECG data analysis device extracts the first main component (PC1) and the second main component (PC2) from the ECG characteristic data and displays the first data (DATA1), second data (DATA2), and third data (DATA3) in graph (b). can be displayed. Graph (b) may be a two-dimensional graph with the first principal component (PC1) and the second principal component (PC2) as variables. The ECG data analysis device may set a first region (MI) where the first disease (eg, myocardial infarction) occurs with a high probability through statistical analysis or a preset algorithm. The user may determine that the third data (DATA3) in graph (b) has a high probability of corresponding to the first disease. By using ECG data with reduced dimensions, the ECG data analysis device can visually express the position of the ECG and its relationship with other ECGs in virtual space.

Meanwhile, the ECG data analysis device may select the number of main components in consideration of data characteristic preservation rate or data operation speed during the analysis of ECG characteristic data. The ECG data analysis device can increase the number of main components to increase the accuracy of the analysis results, or decrease the number of main components to speed up the analysis. The ECG data analysis device may select the number of main components so that the accumulated variance ratio of the data is more than a preset ratio (for example, 90%).

Referring to FIG. 8, the second neural network model 420 may use ECG main component data and disease information determined based on the ECG main component data as learning data. The second neural network model 420 for predicting the user's disease may generate disease prediction data 330 by inputting the ECG main component data 410.

For a description of the second neural network model 420, refer to the description of the neural network model 320 described in FIG. 4.

Referring to FIGS. 2 and 9 , the computing device 100 according to an embodiment of the present disclosure may acquire an electrocardiogram signal (S110). The computing device 100 may be connected to an ECG measurement device and use a measured ECG signal or an ECG signal transmitted through a network.

The computing device 100 may generate ECG feature data by extracting features from the acquired ECG signal (S120). The computing device 100 may extract n features (n is a natural number) based on a preset feature extraction algorithm. Referring to Figure 1, the features extracted from the ECG signal are 'PR INTERVAL', which is the interval from the start point of the P wave to the start point of the QRS wave, 'QRS duration', which is the interval of the QRS wave, and 'QRS duration', which is the area of the + region among the P waves. It can include various features such as 'P AREA' and 'P AMP', which is the voltage value at the point with the highest + value among P waves.

The computing device 100 may infer n features by inputting the ECG signal into a pre-trained first neural network model. For example, the first neural network model may be learned based on the ECG signal and a plurality of features extracted from the ECG signal using a preset algorithm.

The computing device 100 may reduce the dimension of ECG feature data and generate ECG main component data based on principal component analysis (S130). For example, the computing device 100 may reduce the dimension of ECG feature data and generate ECG main component data using the principal component analysis described in FIG. 5 . The principal component may include a first principal component corresponding to the dimensional axis with the largest variance of the ECG principal component data, and more principal components may be extracted according to the ECG feature data.

In an additional embodiment, the computing device 100 uses independent component analysis (ICA), multidimensional scaling, etc. in addition to principal component analysis to reduce the dimension of the ECG characteristic data. The number can be reduced.

The computing device 100 provides ECG feature data with reduced dimensions, thereby making it easier to analyze the ECG feature data.

Referring to FIGS. 2 and 10 , the computing device 100 according to an embodiment of the present disclosure may reduce the dimension of ECG feature data and generate ECG main component data based on principal component analysis (S210). For the process of generating ECG main component data, refer to the description in FIG. 9.

The computing device 100 can visualize ECG main component data (S220). For example, the result of visualizing ECG main component data in the computing device 100 may be graph (a) or graph (b) shown in FIG. 7 . The computing device 100 extracts one main component from the ECG feature data and shows it in a one-dimensional graph, extracts two main components from the ECG feature data and shows them in a two-dimensional graph, or extracts three main components from the ECG feature data and shows them in a two-dimensional graph. It can be shown on a 3D graph.

The computing device 100 can predict a disease corresponding to ECG main component data (S230). The computing device 100 may predict a disease corresponding to ECG main component data using a statistical method. Referring to FIG. 7, when the ECG main component data corresponding to the first disease is located with a high probability in the first area (MI), if the user's ECG data is located in the first area (MI), it is determined to be the first disease. You can. In an additional embodiment, the computing device 100 may train a second neural network model using ECG main component data and disease information determined based on the ECG main component data as learning data. The second neural network model for predicting the user's disease can generate disease prediction data by inputting electrocardiogram main component data.

The computing device 100 may visualize dimensionally reduced ECG feature data and provide it to a user. Users can intuitively understand health status and changes in health status through visualized data.

The various embodiments of the present disclosure described above may be combined with additional embodiments and may be changed within the scope understandable to those skilled in the art in light of the above detailed description. The embodiments of the present disclosure should be understood in all respects as illustrative and not restrictive. For example, each component described as unitary may be implemented in a distributed manner, and similarly, components described as distributed may also be implemented in a combined form. Accordingly, all changes or modified forms derived from the meaning and scope of the claims of the present disclosure and their equivalent concepts should be construed as being included in the scope of the present disclosure.

Claims

An electrocardiogram data analysis method performed by a computing device including at least one processor, comprising:

Acquiring an electrocardiogram signal;

generating electrocardiogram feature data including a plurality of features based on the electrocardiogram signal; and

Reducing the dimension of the ECG feature data based on one or more main components to generate ECG main component data,

The one or more main components are characterized in that the plurality of features are merged based on the correlation between the plurality of features,

method.
According to claim 1,

The step of generating the ECG feature data includes extracting n features (n is a natural number) based on a preset feature extraction algorithm.

method.
According to claim 1,

The step of generating the ECG feature data includes inferring n features by inputting the ECG signal into a pre-trained first neural network model,

method.
According to claim 1,

Characterized in that the principal component includes a first principal component corresponding to the dimensional axis in which the variance of the electrocardiogram principal component data is greatest,

method.
According to claim 4,

Further comprising visualizing the first principal component and the ECG principal component data in a first graph,

method.
According to claim 5,

The first graph is characterized in that it includes first disease area information,

method.
According to claim 4,

Characterized in that the main component further includes a second main component orthogonal to the first main component,

method.
According to claim 8,

Further comprising visualizing the ECG feature data in a second graph based on the first principal component and the second principal component,

method.
According to claim 1,

Inputting the electrocardiogram principal component data into a pre-trained second neural network model to predict a disease corresponding to the electrocardiogram principal component data; further comprising:

method.
A computer program stored in a computer-readable storage medium, wherein the computer program, when executed on one or more processors, performs operations for analyzing electrocardiogram data,

The above operations are:

An operation to obtain an electrocardiogram signal;

generating electrocardiogram feature data including a plurality of features based on the electrocardiogram signal; and

Reducing the dimension of the ECG feature data based on one or more main components to generate ECG main component data;

Including,

The one or more main components are characterized in that the plurality of features are merged based on the correlation between the plurality of features,

computer program.
A computing device for analyzing electrocardiogram data, comprising:

A processor including at least one core;

a memory containing program codes executable on the processor; and

A network unit for acquiring electrocardiogram data;

Including,

The processor,

Obtaining an electrocardiogram signal, generating electrocardiogram feature data including a plurality of features based on the electrocardiogram signal, and reducing the dimension of the electrocardiogram feature data based on one or more principal components to generate electrocardiogram principal component data,

The one or more main components are characterized in that the plurality of features are merged based on the correlation between the plurality of features,

Device.