System for generating electrocardiogram on basis of deep learning algorithm and method thereof

Abstract

To address the difficulty of measuring a 12-lead electrocardiogram at home or in everyday life, since the chest must be exposed to attach chest electrodes and it is difficult for the general public to correctly attach nine electrodes to accurate positions.SOLUTION: The present invention relates to a system for generating electrocardiograms based on a deep learning algorithm and a method thereof. According to the invention, the system comprises: a data input unit that receives 12-lead electrocardiograms measured from a plurality of patients; a data extraction unit that extracts training data from the input 12-lead electrocardiograms; a training unit that inputs the extracted training data into a plurality of training models to train the models on characteristics of the electrocardiograms; an electrocardiogram generation unit that receives one or more reference electrocardiograms from a subject to be measured, and generates virtual electrocardiograms by inputting the reference electrocardiogram into the plurality of trained training models; and a control unit that synchronizes the reference electrocardiogram and the generated virtual electrocardiograms with each other, and outputs waveforms for the synchronized reference electrocardiogram and virtual electrocardiograms.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to an electrocardiogram generation system and method based on a deep learning algorithm, and more particularly to an electrocardiogram generation system and method that generates multiple electrocardiograms from one or more lead electrocardiograms using a deep learning algorithm. [Background technology]

[0002] The standard 12-lead electrocardiogram used in hospitals involves attaching six electrodes to the entire chest and three electrodes to each limb (four electrodes if the ground electrode is included), and then collecting all 12 lead information and combining it to diagnose illnesses.

[0003] A 12-lead electrocardiogram records the cardiac electrical potential in 12 electrical directions centered on the heart, allowing the condition of the heart to be determined in multiple directions, making it possible to accurately diagnose cardiac diseases that are limited to a single area.

[0004] Measuring the cardiac potential in multiple directions is significant in that it allows the characteristics of the heart to be understood in each direction, and for this reason, medical professionals recommend measuring the standard 12-lead electrocardiogram to diagnose cardiac diseases (such as myocardial infarction).

[0005] However, in order to take 12-lead images, the chest must be exposed to attach chest electrodes, and it is difficult for an average person to attach nine electrodes (three on the limbs and six on the chest) in the correct positions, making it difficult to measure at home or in daily life. Also, it is difficult to move after attaching the 10 electrodes, making it difficult to use for real-time monitoring.

[0006] Therefore, recently, devices capable of measuring one-lead electrocardiograms or two or more leads have been developed so that they can be used in everyday life.

[0007] First, one-lead ECG devices using two electrodes include watch-type ECG devices (Apple Watch or Galaxy Watch). In watch-type ECG devices, the back of the watch touches the left wrist, and the right finger touches the crown of the watch, connecting the left arm electrode and the right arm electrode. The potential difference between the two electrodes is used to measure the I-lead ECG.

[0008] The watch-type electrocardiogram mechanism is worn on the left arm, and lead I is measured by touching the crown with the right hand, lead II is measured by touching the crown with the right hand while the watch is placed on the abdomen, and lead III is measured by touching the crown with the left hand while the watch is placed on the abdomen.Then, a V1-6 lead electrocardiogram is measured by touching the back of the watch to the V1-6 electrode position while touching the watch crown with the left hand.

[0009] The above method has the drawback of being less user-friendly since the user must accurately contact the V1-6 position for the electrocardiogram. In addition, the V1-6 lead electrocardiogram cannot implement standard chest leads because it illustrates the potential difference between the right arm electrode and the V electrode, unlike a standard chest lead electrocardiogram which must illustrate the potential difference between a virtual center point and the V electrode even when the user contacts the relevant area.

[0010] Alternatively, two or more lead ECGs can be measured by holding the electrodes in both hands, contacting the electrodes on the back of the device to the leg or ankle, and contacting electrodes on the left arm, right arm, and left leg, respectively. However, this method has the disadvantage that it cannot be used for long-term monitoring (1 hour, 24 hours, 7 days, etc.), and the device itself must have three electrodes.

[0011] The background technology of the present invention is disclosed in Korean Patent Registration No. 10-2180135 (published on November 17, 2020). Summary of the Invention [Problem to be solved by the invention]

[0012] Thus, the present invention provides an electrocardiogram generation system and method that uses a deep learning algorithm to generate multiple electrocardiograms from one or more lead electrocardiograms. [Means for solving the problem]

[0013] According to an embodiment of the present invention for achieving this technical objective, an electrocardiogram generation system based on a deep learning algorithm includes: a data input unit that receives 12-lead electrocardiograms measured on a plurality of patients and patient information corresponding to the 12-lead electrocardiograms; a data extraction unit that classifies and stores the input 12-lead electrocardiograms according to the patient information and extracts learning data from the stored 12-lead electrocardiograms; a learning unit that inputs the extracted learning data into one or more learning models to learn characteristics of the electrocardiogram; an electrocardiogram generation unit that receives one or more reference electrocardiograms from the subject and inputs the input reference electrocardiograms into one or more learning models that have completed learning to generate a virtual electrocardiogram; and a control unit that synchronizes the reference electrocardiogram and the generated virtual electrocardiogram with each other and outputs waveforms for the synchronized reference electrocardiogram and virtual electrocardiogram.

[0014] The patient information may include at least one of gender, age, presence or absence of heart disease, and a potential vector of a measured electrocardiogram.

[0015] The learning unit inputs an n-lead electrocardiogram, which is a portion of a limb 6-lead electrocardiogram and a chest 6-lead electrocardiogram, into a first learning model, thereby training the first learning model to discriminate potential vectors for the input electrocardiogram and generate 12-n electrocardiograms.

[0016] The learning unit constructs 12 second learning models for the limb 6-lead electrocardiogram and the chest 6-lead electrocardiogram, and when 12-n generated electrocardiograms output from the constructed first learning model are input to the second learning model, the second learning model can be trained to convert the input lead electrocardiograms into a style having corresponding potential vectors and output them as n virtual electrocardiograms.

[0017] The first and second learning models may be constructed by combining a generative adversarial network and an autoencoder method.

[0018] The electrocardiogram generation unit inputs the input reference electrocardiogram into a first learning model, extracts potential vectors for the reference electrocardiogram to generate the remaining multiple virtual electrocardiograms, and inputs the generated virtual electrocardiogram into a second learning model trained using the extracted potential vectors, thereby again generating a virtual electrocardiogram with leads identical to the reference electrocardiogram.

[0019] The control unit matches and synchronizes the reference electrocardiogram with the plurality of virtual electrocardiograms, and outputs the synchronized plurality of electrocardiograms via a monitor.

[0020] According to another embodiment of the present invention, an electrocardiogram generating method using an electrocardiogram generating system includes the steps of: receiving 12-lead electrocardiograms measured on a plurality of patients and patient information corresponding to the 12-lead electrocardiograms; classifying and storing the input 12-lead electrocardiograms according to the patient information; extracting learning data from the stored 12-lead electrocardiograms; inputting the extracted learning data into one or more learning models to learn characteristics of the electrocardiograms; receiving one or more reference electrocardiograms from the subjects and inputting the input reference electrocardiograms into one or more learning models that have completed learning to generate a virtual electrocardiogram; and synchronizing the reference electrocardiogram and the generated virtual electrocardiogram with each other and outputting waveforms for the synchronized reference electrocardiogram and virtual electrocardiogram. [Effects of the Invention]

[0021] As described above, according to the present invention, multiple synchronized electrocardiograms can be generated from two electrocardiograms measured at different points using a deep learning algorithm, thereby enabling accurate diagnosis of cardiac diseases, just like a 12-lead electrocardiogram.

[0022] In addition, according to the present invention, since a two-lead electrocardiogram is used, it can be measured at home or in daily life, and can also be measured while moving, allowing for real-time monitoring. Furthermore, since a synchronized electrocardiogram is generated using electrocardiogram information measured at different times, medical staff can interpret the synchronized lead electrocardiogram information according to the corresponding heartbeat, allowing for more accurate diagnosis. [Brief explanation of the drawings]

[0023] [Figure 1] 1 is a configuration diagram for explaining an electrocardiogram generating system according to an embodiment of the present invention. [Figure 2] 1 is a flowchart illustrating an electrocardiogram generating method using an electrocardiogram generating system according to an embodiment of the present invention. [Figure 3] FIG. 1 is an illustrative diagram illustrating a method for measuring a general 12-lead electrocardiogram. [Figure 4] FIG. 3 is an exemplary diagram illustrating step S240 shown in FIG. 2. [Figure 5] FIG. 3 is an exemplary diagram illustrating step S260 shown in FIG. 2. DETAILED DESCRIPTION OF THE INVENTION

[0024] The electrocardiogram generating system 100 according to the embodiment of the present invention includes a data input unit 110, a data extraction unit 120, a learning unit 130, an electrocardiogram generating unit 140, and a control unit 150.

[0025] First, the data input unit 110 receives 12-lead electrocardiograms measured on a plurality of patients and patient information corresponding to the 12-lead electrocardiograms.

[0026] Here, the 12-lead electrocardiogram includes a 4-lead electrocardiogram and a 6-lead chest electrocardiogram, and the patient information includes at least one of gender, age, presence or absence of heart disease, and the potential vector of the measured electrocardiogram.

[0027] The data extraction unit 120 classifies the 12-lead electrocardiograms using the input patient information and stores them in a database. The data extraction unit 120 then randomly extracts multiple 12-lead electrocardiograms stored in the database to generate training data.

[0028] The learning unit 130 then trains the constructed learning model using the training data. More specifically, the learning unit 130 constructs a first learning model that extracts potential vector information from the input lead electrocardiogram and generates a 12-n-lead virtual electrocardiogram from the n-lead reference electrocardiogram, and a second learning model that generates an n-lead virtual electrocardiogram from the 12-n-lead virtual electrocardiogram.

[0029] Then, the learning unit 130 inputs the generated learning data into the constructed first learning model or second learning model, and trains the first learning model or second learning model.

[0030] The electrocardiogram generation unit 140 acquires one or more reference electrocardiograms from the subject. Then, the electrocardiogram generation unit 140 inputs the acquired reference electrocardiograms into a first learning model that has completed learning, thereby generating a first virtual electrocardiogram based on potential vectors for the reference electrocardiogram. The electrocardiogram generation unit 140 inputs the generated first virtual electrocardiogram into a trained second learning model, thereby causing the second learning model to output a second virtual electrocardiogram.

[0031] Finally, the control unit 150 synchronizes the reference electrocardiogram with the virtual electrocardiogram, and outputs a plurality of synchronized electrocardiograms from the reference electrocardiogram and the virtual electrocardiogram via the monitor device.

[0032] MODE FOR CARRYING OUT THE INVENTION

[0033] Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings, in which the thickness of lines and the size of components shown may be exaggerated for clarity and convenience of explanation.

[0034] In addition, the terms described below are defined in consideration of their functions in the present invention, and may vary depending on the intentions or practices of users or operators. Therefore, the definitions of such terms should be based on the overall content of this specification.

[0035] The following describes in more detail the deep learning algorithm-based electrocardiogram generating system according to an embodiment of the present invention with reference to FIG.

[0036] FIG. 1 is a configuration diagram for explaining an electrocardiogram generating system according to an embodiment of the present invention.

[0037] As shown in FIG. 1, an electrocardiogram generating system 100 according to an embodiment of the present invention includes a data input unit 110, a data extraction unit 120, a learning unit 130, an electrocardiogram generating unit 140, and a control unit 150.

[0038] First, the data input unit 110 receives 12-lead electrocardiograms measured on a plurality of patients and patient information corresponding to the 12-lead electrocardiograms.

[0039] Here, the 12-lead electrocardiogram includes a 4-lead electrocardiogram and a 6-lead chest electrocardiogram, and the patient information includes at least one of gender, age, presence or absence of heart disease, and the potential vector of the measured electrocardiogram.

[0040] The data extraction unit 120 classifies the 12-lead electrocardiograms using the input patient information and stores them in a database. The data extraction unit 120 then randomly extracts multiple 12-lead electrocardiograms stored in the database to generate training data.

[0041] The learning unit 130 then uses the learning data to train the constructed learning model. More specifically, the learning unit 130 extracts potential vector information from the input lead electrocardiogram and constructs a first learning model that generates a 12-n-lead virtual electrocardiogram from the n-lead reference electrocardiogram, and a second learning model that generates an n-lead virtual electrocardiogram from the 12-n-lead virtual electrocardiogram.

[0042] Then, the learning unit 130 inputs the generated learning data into the constructed first learning model or second learning model, and trains the first learning model or second learning model.

[0043] The electrocardiogram generation unit 140 acquires one or more reference electrocardiograms from the subject. Then, the electrocardiogram generation unit 140 inputs the acquired reference electrocardiograms into a first learning model that has completed learning, thereby generating a first virtual electrocardiogram based on potential vectors for the reference electrocardiogram. The electrocardiogram generation unit 140 inputs the generated first virtual electrocardiogram into a trained second learning model, thereby causing the second learning model to output a second virtual electrocardiogram.

[0044] Finally, the control unit 150 synchronizes the reference electrocardiogram with the virtual electrocardiogram, and outputs a plurality of synchronized electrocardiograms from the reference electrocardiogram and the virtual electrocardiogram via the monitor device.

[0045] An electrocardiogram generating method using the electrocardiogram generating system 100 according to an embodiment of the present invention will be described in more detail below with reference to FIGS.

[0046] FIG. 2 is a flowchart illustrating a method for generating an electrocardiogram using the electrocardiogram generating system according to an embodiment of the present invention.

[0047] As shown in FIG. 2, the method for generating an electrocardiogram using the electrocardiogram generating system according to an embodiment of the present invention is divided into a step of training a learning model and a step of generating an electrocardiogram using the trained learning model.

[0048] In the stage of training the learning model, first, lead electrocardiograms measured on a plurality of patients and patient information are input to the electrocardiogram generation system 100 (S210).

[0049] First, a 12-lead electrocardiogram records the cardiac potential in 12 electrical directions centered on the heart.

[0050] The general method for taking an electrocardiogram is to first attach four electrodes (limb leads) to the patient's arms and legs. The electrocardiogram device measures the potential using the limb leads and the left leg lead, while the right leg lead serves as a ground electrode.

[0051] FIG. 3 is an illustrative diagram for explaining a method for measuring a general 12-lead electrocardiogram.

[0052] As shown in Figure 3, the electrocardiogram measurement device generates a lead I electrocardiogram by subtracting the potential of the right arm from the potential of the left arm, generates a lead II electrocardiogram by subtracting the potential of the right arm from the potential of the left leg, and generates a lead III electrocardiogram by subtracting the potential of the left arm from the potential of the left leg.

[0053] The electrocardiogram measuring device calculates the potential of the virtual center point by averaging the potentials of both arm and left leg electrodes. The electrocardiogram measuring device then generates an electrocardiogram for lead aVL by subtracting the potential of the virtual center point from the potential of the left arm, and generates an electrocardiogram for lead aVR by subtracting the potential of the virtual center point from the potential of the right arm. The electrocardiogram measuring device also generates an electrocardiogram for lead aVF by subtracting the potential of the virtual center point from the potential of the left leg electrode. As described above, the electrocardiogram measuring device generates a total of six leads (six columns) of electrocardiograms using three limb electrodes (four including the ground electrode).

[0054] The electrocardiogram measurement device then generates a six-lead chest electrocardiogram using the difference between the potential of the virtual center point determined by the limb leads and the potential of six electrodes attached to the chest. Six electrodes, V1, V2, V3, V4, V5, and V6, are attached to the chest from the front to the left side of the chest at predetermined positions.

[0055] The electrocardiogram measurement device then generates a V1 lead electrocardiogram by subtracting the potential measured at the V1 electrode from the potential at the virtual center point obtained by averaging the potential of the four limb electrodes.

[0056] As described above, the electrocardiogram measuring device generates a 12-lead electrocardiogram using a plurality of electrodes attached to the patient, and transmits the generated 12-lead electrocardiogram to the electrocardiogram generating system 100.

[0057] Here, the electrocardiogram generating system 100 further receives the generated 12-lead electrocardiogram and corresponding patient information.

[0058] Here, the patient information includes at least one of the following: sex, age, presence or absence of heart disease, and potential vectors of the measured electrocardiogram.

[0059] After step S210 is completed, the electrocardiogram generating system 100 extracts learning data using the collected 12-lead electrocardiogram and patient information (S220).

[0060] To further explain this, the data extraction unit 120 classifies the collected 12-lead ECGs according to patient information and stores them in a database. The data extraction unit 120 then randomly extracts data from the stored 12-lead ECGs to generate training data.

[0061] Thereafter, the learning unit 130 uses the generated learning data to train the first learning model and the second learning model (S230).

[0062] First, the learning unit 130 inputs learning data composed of a 12-lead electrocardiogram into the first learning model and the second learning model to learn the characteristics of the electrocardiogram. To explain this further, the direction of electrical flow in a lead electrocardiogram varies depending on the potential vector, and the patient's age and gender affect the electrocardiogram style. That is, as people get older, the cardiac muscle decreases, which tends to decrease the amplitude of the electrocardiogram. In women, the position of the electrocardiogram electrodes is lowered due to the presence of breasts, or the distance between the heart and the electrodes increases, resulting in deformation of the electrocardiogram.

[0063] In addition, if there is chronic lung disease (late obstructive pulmonary disease), the lung capacity increases and the heart, which is located between the lungs, is forced to stand vertically, causing the electrical flow of the heart to change vertically in three-dimensional space.

[0064] Therefore, the learning unit 130 inputs the lead electrocardiograms, which are separated and stored according to patient information, i.e., the patient's age, sex, health level, and potential vectors, into the first and second learning models. The first and second learning models then learn the characteristics of the input lead electrocardiograms and extract potential vector information from the input lead electrocardiograms. However, it is also possible to receive and learn an electrocardiogram without separating it according to patient information, and in this manner, it is also possible to train a learning model that is not limited to one characteristic.

[0065] The learning unit 130 also constructs 12 second learning models corresponding to the 12-lead electrocardiogram, and trains each second learning model according to the characteristics of the electrocardiogram.

[0066] To further explain this, the first learning model learns the correlation between the input electrocardiogram and the 12-lead electrocardiogram based on a deep learning algorithm, extracts at least one of the features of the input electrocardiogram, i.e., age, sex, and potential vector, and thereby generates a 12-n-lead virtual electrocardiogram from the n-lead reference electrocardiogram, while the second learning model generates an n-lead virtual electrocardiogram from the 12-n-lead virtual electrocardiogram. For example, assuming that a V1 lead electrocardiogram is to be generated, the learning unit 130 trains a style in the first learning model to generate a virtual lead I, II, III, aVL, aVR, aVF, V2, V3, V4, V5, and V6 electrocardiogram from a lead V1 electrocardiogram, and trains a style in the second learning model to generate a virtual lead V1 electrocardiogram from a virtual lead I, II, III, aVL, aVR, aVF, V2, V3, V4, V5, and V6 electrocardiogram.The second learning model then converts the input electrocardiogram into a V1 lead style to generate a virtual electrocardiogram.

[0067] The first learning model is trained by inputting the lead of the input electrocardiogram so that it can determine which lead the input electrocardiogram belongs to. As a result, when used after training, it can generate a 12-lead electrocardiogram even if an arbitrary electrocardiogram whose leads are unknown is input.

[0068] Here, the first learning model and the second learning model are based on a deep learning algorithm composed of an autoencoder or a generative adversarial network, and the deep learning algorithm may be implemented using one selected from the autoencoder or the generative adversarial network, or may be implemented by combining the autoencoder and the generative adversarial network.

[0069] When the learning of the learning model is completed in steps S210 and S230, the electrocardiogram generating system 100 according to an embodiment of the present invention generates an electrocardiogram using the learned learning model.

[0070] First, the electrocardiogram generating system 100 receives a reference electrocardiogram measured through electrodes attached to the body of a subject (S240).

[0071] Here, the reference electrocardiogram does not contain information about the potential vectors.

[0072] Thereafter, the electrocardiogram generating unit 140 inputs the input reference electrocardiogram into the first learning model and the second learning model to generate a plurality of virtual electrocardiograms (S250).

[0073] First, the electrocardiogram generating unit 140 inputs a reference electrocardiogram into the first learning model to extract characteristics of the reference electrocardiogram.

[0074] Here, the characteristics include at least one of the age, sex, and potential vector of the subject.

[0075] As a result, a first virtual electrocardiogram is generated from the reference electrocardiogram.

[0076] Then, the electrocardiogram generation unit 140 inputs the generated first virtual electrocardiogram into the second learning model, which then generates a second virtual electrocardiogram based on the input first virtual electrocardiogram.

[0077] FIG. 4 is an example diagram for explaining step S240 shown in FIG.

[0078] As shown in FIG. 4, the first electrocardiogram represents a reference electrocardiogram. Assume that the first learning model has characteristics of the reference electrocardiogram as L1. Then, the electrocardiogram generation unit 140 generates virtual electrocardiograms of the remaining 11 leads based on the L1 electrocardiogram. The generated virtual electrocardiogram of 11 leads is input to a second learning model, which then generates a virtual L1 electrocardiogram. Each learning model includes a generator that generates an electrocardiogram and a discriminator that determines whether the generated electrocardiogram is accurate and uses feedback to improve the accuracy of the generated electrocardiogram.

[0079] After completing step S250, the control unit 150 outputs the reference electrocardiogram and 11 virtual electrocardiograms through the monitoring device (S260).

[0080] FIG. 5 is an example diagram for explaining step S260 shown in FIG.

[0081] The control unit 150 outputs a reference electrocardiogram and 11 virtual electrocardiograms. Then, as shown in Fig. 5, the control unit 150 synchronizes the output reference electrocardiogram with the 11 virtual electrocardiograms to determine whether or not the subject has a health abnormality. Here, since a virtual electrocardiogram of the same leads as the reference electrocardiogram is also generated by the second learning model, it is also possible to make a determination using the 12 virtual electrocardiograms.

[0082] As described above, the slope and amplitude of the electrocardiogram output vary depending on the age, sex, and health condition of the subject. Therefore, the control unit 150 outputs the reference electrocardiogram and 12 electrocardiograms synchronized with the 11 virtual electrocardiograms via the monitoring device.

[0083] As described above, according to the present invention, multiple synchronized electrocardiograms can be generated from two electrocardiograms measured at different points using a deep learning algorithm, thereby enabling accurate interpretation of cardiac diseases, just like a 12-lead electrocardiogram.

[0084] In addition, the electrocardiogram generating system according to the present invention uses two-lead electrocardiograms, so that measurements can be taken at home or in daily life, even while moving, enabling real-time monitoring. Furthermore, a synchronized electrocardiogram is generated using electrocardiogram information measured at different times, so that medical staff can interpret the synchronized lead electrocardiogram information according to the corresponding heartbeat, enabling more accurate diagnosis.

[0085] Although the present invention has been described based on the embodiments shown in the drawings, these are merely illustrative, and those skilled in the art will recognize that various modifications and equivalent embodiments are possible. Therefore, the true technical scope of the present invention should be determined by the technical spirit of the following claims. [Industrial Applicability]

[0086] The present invention can generate multiple synchronized electrocardiograms from two electrocardiograms measured at different points using a deep learning algorithm, allowing accurate interpretation of cardiac diseases like a 12-lead electrocardiogram, and can be industrially applied to electrocardiogram generation systems based on various deep learning algorithms. [Explanation of symbols]

[0087] 100 Electrocardiogram Generation System 110 Data Entry Section 120 Data Extraction Unit 130 Learning Department 140 Electrocardiogram Generation Unit 150 control section

Claims

1. An electrocardiogram generation system based on a deep learning algorithm, comprising: a data input unit for inputting 12-lead electrocardiograms measured on a plurality of patients; a data extraction unit that extracts learning data from the input 12-lead electrocardiogram; a learning unit that inputs the extracted learning data into a plurality of learning models to learn electrocardiogram characteristics; an electrocardiogram generating unit that receives one or more reference electrocardiograms from the measured subject and inputs the input reference electrocardiograms into a plurality of learning models that have completed learning to generate a virtual electrocardiogram; a control unit that synchronizes the reference electrocardiogram and the generated virtual electrocardiogram with each other and outputs waveforms corresponding to the synchronized reference electrocardiogram and virtual electrocardiogram. Electrocardiogram generation system.

2. The patient information includes at least one of gender, age, presence or absence of heart disease, and a potential vector of a measured electrocardiogram. The electrocardiogram generating system according to claim 1 .

3. the learning unit inputs a four-limb six-lead electrocardiogram and a chest six-lead electrocardiogram into a first learning model, and causes the first learning model to learn to discriminate potential vectors corresponding to the input electrocardiograms; The electrocardiogram generating system according to claim 2 .

4. the learning unit constructs a first learning model and a second learning model for a limb six-lead electrocardiogram and a chest six-lead electrocardiogram, and when a reference lead electrocardiogram is input to the constructed first learning model and second learning model, the first learning model and the second learning model are trained to convert the input reference lead electrocardiogram into a style having a corresponding potential vector and output it as a virtual electrocardiogram. The electrocardiogram generating system according to claim 3 .

5. The first learning model and the second learning model are constructed by mixing or respectively using a generative adversarial network and an autoencoder method.

5. The electrocardiogram generating system according to claim 3 or 4.

6. the electrocardiogram generation unit inputs the input reference electrocardiogram into a first learning model to extract a potential vector for the reference electrocardiogram, generates a first virtual electrocardiogram, and inputs the first virtual electrocardiogram into a second learning model to generate a second virtual electrocardiogram; The electrocardiogram generating system according to claim 5 .

7. the control unit matches and synchronizes the reference electrocardiogram with the plurality of virtual electrocardiograms, outputs the synchronized plurality of electrocardiograms via a monitor, and determines whether or not the subject has a health abnormality using at least one of amplitude, gradient, and electrode position of the waveforms of the output reference electrocardiogram and virtual electrocardiogram. The electrocardiogram generating system according to claim 6 .

8. An electrocardiogram generating method using an electrocardiogram generating system, comprising: receiving 12-lead electrocardiograms measured on a plurality of patients and patient information corresponding to the 12-lead electrocardiograms; classifying and storing the input 12-lead electrocardiograms according to the patient information, and extracting learning data from the stored 12-lead electrocardiograms; inputting the extracted learning data into a plurality of learning models to learn electrocardiogram characteristics; receiving one or more reference electrocardiograms from the measured subject, and inputting the input reference electrocardiograms into a plurality of learning models that have completed learning to generate a virtual electrocardiogram; and synchronizing the reference electrocardiogram and the generated virtual electrocardiogram with each other, and outputting waveforms of the synchronized reference electrocardiogram and virtual electrocardiogram and a result indicating whether or not a health abnormality has occurred in the measured subject. Electrocardiogram generation method.

9. The patient information includes at least one of gender, age, presence or absence of heart disease, and a potential vector of a measured electrocardiogram. The method for generating an electrocardiogram according to claim 8.

10. The step of learning the characteristics of the electrocardiogram includes inputting a four-limb six-lead electrocardiogram and a chest six-lead electrocardiogram into a first learning model, and training the first learning model to discriminate potential vectors for the input electrocardiograms and generate a first virtual electrocardiogram. The method for generating an electrocardiogram according to claim 8.

11. The step of learning the characteristics of the electrocardiogram includes constructing a second learning model that learns a style based on the first virtual electrocardiogram, and training the constructed second learning model so that when the first virtual electrocardiogram is input to the constructed second learning model, the input lead electrocardiogram is converted into a style having a corresponding potential vector and output as a second virtual electrocardiogram. The method for generating an electrocardiogram according to claim 10.

12. The first learning model and the second learning model are constructed by mixing or respectively using a generative adversarial network and an autoencoder method.

12. The method for generating an electrocardiogram according to claim 10 or 11.

13. The step of generating the virtual electrocardiogram includes inputting the input reference electrocardiogram into a first learning model to extract potential vectors for the reference electrocardiogram and generate a first virtual electrocardiogram, and inputting the first virtual electrocardiogram into a second learning model trained using the extracted potential vectors to generate a second virtual electrocardiogram. The method for generating an electrocardiogram according to claim 12.

14. The step of outputting the presence or absence of a health abnormality includes matching and synchronizing the reference electrocardiogram and the plurality of virtual electrocardiograms, outputting the synchronized plurality of electrocardiograms via a monitor, and determining the presence or absence of a health abnormality of the subject using at least one of amplitude, gradient, and electrode position of the waveforms of the output reference electrocardiogram and the virtual electrocardiogram. The method for generating an electrocardiogram according to claim 13.

Images