Method, system and non-transitory computer-readable recording medium for managing training data for a biological signal analysis model

Abstract

According to one aspect of the present invention, there is provided a method for managing training data for a biological signal analysis model, the method including a step of converting data relating to at least one lead of a plurality of leads into augmented data relating to a specific lead of the plurality of leads, and a step of using the augmented data and the data relating to the specific lead as training data to train an analysis model for discriminating arrhythmias using the data relating to the specific lead.

Description

[Technical field]

[0001] The present invention relates to a method, a system, and a non-transitory computer-readable recording medium for managing training data for a biological signal analysis model. [Background technology]

[0002] Recently, the quality of life of all human beings has improved due to the rapid development of science and technology, and many changes have been made in the medical environment. In particular, wearable monitoring devices that can analyze electrocardiograms and detect arrhythmias in daily life without going to the hospital have become widely used by the masses.

[0003] Generally, a widely used method for detecting arrhythmia is to analyze a 12-lead ECG. In order to analyze a 12-lead ECG, multiple contacts must be formed on the subject's body, but the characteristics of wearable monitoring devices generally limit the ability to form multiple contacts on the subject's body, so arrhythmia is detected by analyzing only data related to specific leads of the 12-lead ECG.

[0004] Such wearable monitoring devices will be equipped with artificial intelligence models to analyze data related to specific leads, but traditionally, the training data used to train the artificial intelligence models has only used data related to specific leads.

[0005] However, when the artificial intelligence model is trained using only data related to a specific lead, while it is possible to make highly accurate judgments about some types of arrhythmias that can be accurately identified by analyzing only the data related to the specific lead, there is a problem in that it is difficult to make highly accurate judgments about types of arrhythmias that cannot be accurately identified without a comprehensive analysis of the data related to the specific lead and the data related to other leads. Summary of the Invention [Problem to be solved by the invention]

[0006] An object of the present invention is to solve all of the problems of the prior art described above.

[0007] Another object of the present invention is to enable an artificial intelligence model for discriminating arrhythmias using data related to a specific lead to be trained using data related to multiple leads, thereby enabling the artificial intelligence model to discriminate various types of arrhythmias with high accuracy. [Means for solving the problem]

[0008] A representative configuration of the present invention for achieving the above object is as follows.

[0009] According to one aspect of the present invention, there is provided a method for managing training data for a biological signal analysis model, the method including a step of converting data relating to at least one lead of a plurality of leads into augmented data relating to a specific lead of the plurality of leads, and a step of using the augmented data and the data relating to the specific lead as training data to train an analysis model for discriminating arrhythmias using the data relating to the specific lead.

[0010] According to another aspect of the present invention, there is provided a system for managing learning data for a biological signal analysis model, the system including a data management unit that converts data relating to at least one of a plurality of leads into augmented data relating to a specific lead of the plurality of leads, and a learning management unit that uses the augmented data and the data relating to the specific lead as learning data to train an analysis model for discriminating arrhythmia using the data relating to the specific lead.

[0011] In addition, other methods for embodying the present invention, other systems, and a non-transitory computer readable recording medium having a computer program for carrying out the methods are also provided. Effect of the Invention

[0012] According to the present invention, an artificial intelligence model for identifying arrhythmias using data related to a specific lead is trained using data related to multiple leads, thereby enabling the artificial intelligence model to identify various types of arrhythmias with high accuracy. [Brief description of the drawings]

[0013] [Figure 1] 1 is a diagram showing a schematic configuration of an entire system for managing learning data of a biosignal analysis model according to an embodiment of the present invention. [Diagram 2] 2 is a diagram illustrating in detail the internal configuration of a learning data management system according to an embodiment of the present invention. [Diagram 3] 1 is a diagram illustrating an example of data related to multiple leads that has been vectorized according to one embodiment of the present invention. [Figure 4] 1 is a diagram illustrating an example of vectorized data regarding other leads and vectorized data regarding a specific lead according to an embodiment of the present invention. [Diagram 5] 1 is an exemplary diagram showing enhanced data associated with a particular lead according to one embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] The following detailed description of the present invention will be made with reference to the accompanying drawings, which show, by way of example, specific embodiments in which the present invention may be practiced. Such embodiments are described in detail to enable those skilled in the art to fully practice the present invention. It should be understood that the various embodiments of the present invention are different from one another but are not necessarily mutually exclusive. For example, specific shapes, structures and characteristics described herein may be embodied in one embodiment without departing from the spirit and scope of the present invention. It should also be understood that the location or arrangement of individual components within each embodiment may be changed without departing from the spirit and scope of the present invention. Therefore, the following detailed description is not to be taken in a limiting sense, and the scope of the present invention should be understood to include the scope of the claims and all equivalents thereto. In the drawings, like reference numerals indicate the same or similar components throughout the various aspects.

[0015] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, various preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings in order to enable those skilled in the art to which the present invention pertains to easily practice the present invention.

[0016] Overall system configuration

[0017] FIG. 1 is a diagram showing a schematic configuration of an entire system for managing learning data of a biosignal analysis model according to an embodiment of the present invention.

[0018] As shown in FIG. 1, the entire system according to an embodiment of the present invention may include a communication network 100, a learning data management system 200, and a device 300.

[0019] First, the communication network 100 according to an embodiment of the present invention may be configured regardless of the type of communication, such as wired communication or wireless communication, and may be configured as various communication networks, such as a local area network (LAN), a metropolitan area network (MAN), and a wide area network (WAN). Preferably, the communication network 100 in this specification may be the well-known Internet or World Wide Web (WWW). However, the communication network 100 is not limited thereto, and may include at least a part of a well-known wired / wireless data communication network, a well-known telephone network, or a well-known wired / wireless television communication network.

[0020] For example, communication network 100 may be a wireless data communication network that embodies, at least in part, conventional communication methods such as radio frequency (RF) communication, WiFi communication, cellular (e.g., LTE) communication, Bluetooth communication (more specifically, Bluetooth Low Energy (BLE)), infrared communication, ultrasonic communication, and the like.

[0021] Next, the learning data management system 200 according to one embodiment of the present invention can perform a function of converting data relating to at least one of a plurality of leads into augmented data relating to a specific lead among the plurality of leads, using the augmented data and the data relating to the specific lead as learning data, and learning an analytical model for discriminating arrhythmia using the data relating to the specific lead.

[0022] The configuration and functions of the learning data management system 200 according to an embodiment of the present invention will be explained in detail below.

[0023] Next, the device 300 according to one embodiment of the present invention is a digital device having a function of being able to communicate after being connected to the learning data management system 200, and any digital device having a memory means, a microprocessor, and computing power, such as a smartphone, tablet, smart watch, smart patch, smart band, smart glasses, desktop computer, notebook computer, workstation, PDA, web pad, mobile phone, etc., may be adopted as the device 300 according to the present invention.

[0024] In particular, the device 300 according to an embodiment of the present invention may include sensing means (eg, contact electrodes, etc.) for acquiring a predetermined biosignal (eg, an electrocardiogram) from the human body.

[0025] Meanwhile, the device 300 according to an embodiment of the present invention may include an application (not shown) that assists the user to receive the service according to the present invention from the learning data management system 200. Such an application may be downloaded from the learning data management system 200 or an external application distribution server (not shown). Meanwhile, the characteristics of such an application may be generally similar to the data management unit 210, the learning management unit 220, the communication unit 230, and the control unit 240 of the learning data management system 200 described below. Here, at least a part of the application may be replaced with a hardware device or a firmware device capable of performing substantially the same or equivalent functions as the application, if necessary.

[0026] Learning Data Management System Configuration

[0027] Hereinafter, the internal configuration of the learning data management system 200, which performs important functions for implementing the present invention, and the functions of each component will be described in detail.

[0028] FIG. 2 is a diagram illustrating in detail the internal configuration of a learning data management system 200 according to an embodiment of the present invention.

[0029] As shown in FIG. 2, the learning data management system 200 according to an embodiment of the present invention may include a data management unit 210, a learning management unit 220, a communication unit 230, and a control unit 240. According to an embodiment of the present invention, the data management unit 210, the learning management unit 220, the communication unit 230, and the control unit 240 of the learning data management system 200 may be program modules, at least some of which communicate with an external system (not shown). Such program modules may be included in the learning data management system 200 in the form of an operating system, application module, or other program modules, and may be physically stored in various known storage devices. In addition, such program modules may be stored in a remote storage device that can communicate with the learning data management system 200. Meanwhile, such program modules include, but are not limited to, routines, subroutines, programs, objects, components, data structures, etc. that perform specific tasks or execute specific abstract data types according to the present invention, which will be described later.

[0030] Meanwhile, although the learning data management system 200 has been described above, such description is illustrative, and it will be apparent to those skilled in the art that at least some of the components or functions of the learning data management system 200 may be embodied within the device 300 or a server (not shown) or included in an external system (not shown) as necessary.

[0031] First, the data management unit 210 according to an embodiment of the present invention may perform a function of converting data related to at least one of a plurality of leads into augmented data related to a particular lead of the plurality of leads.

[0032] The data management unit 210 according to an embodiment of the present invention can acquire data related to a plurality of leads from at least one electrocardiograph (e.g., a 12-lead electrocardiograph, a Holter electrocardiograph, a wearable electrocardiograph, etc.). Here, the data related to a plurality of leads is data related to a 12-lead electrocardiogram (12-lead ECG) and can include data related to leadI, leadII, leadIII, lead aVR, lead aVL, lead aVF, lead V1, lead V2, lead V3, lead V4, lead V5, and lead V6.

[0033] Next, the data management unit 210 according to an embodiment of the present invention can extract data on a specific lead and data on at least one other lead other than the data on the specific lead (hereinafter, referred to as "data on another lead") from data on a plurality of leads acquired from at least one electrocardiograph. Here, the data on the specific lead is learning data used for learning an analysis model for discriminating arrhythmia, and may be data of the same type (i.e., measured on the same lead) as test input data inputted to the analysis model in the analysis process. Also, the data on the other lead is learning data that can be further used together with the data on the specific lead for learning an analysis model for discriminating arrhythmia, and may be data measured on another lead having a predetermined relationship with the data on the specific lead.

[0034] Specifically, the data management unit 210 according to an embodiment of the present invention can create a database of combinations of data to be used as learning data for the analysis model for each test input data input to the analysis model, and based on this, extract data on a specific lead and data on other leads from data on multiple leads acquired from at least one electrocardiograph. For example, when the test input data input to the analysis model is data on lead II, the data management unit 210 according to an embodiment of the present invention can refer to one of the combinations and extract data on lead II as data on a specific lead and data on lead aVR and data on lead I as data on other leads from data on multiple leads acquired from at least one electrocardiograph.

[0035] Meanwhile, the data management unit 210 according to an embodiment of the present invention may convert the extracted data related to other leads into augmented data related to a specific lead. For example, the data management unit 210 according to an embodiment of the present invention may convert the data related to lead aVR and the data related to leadI extracted as data related to other leads into augmented data related to the data related to leadII extracted as data related to a specific lead.

[0036] Specifically, the data management unit 210 according to an embodiment of the present invention can vectorize data related to other leads and data related to a specific lead. More specifically, the data management unit 210 according to an embodiment of the present invention can vectorize data related to other leads and data related to a specific lead in a three-dimensional space. Referring to FIG. 3, data related to leadI, leadII, leadIII, lead aVL, lead aVR, and lead aVF of the 12-lead electrocardiogram can be vectorized based on a yz plane in the three-dimensional space. Also, data related to lead V1, lead V2, lead V3, lead V4, lead V5, and lead V6 of the 12-lead electrocardiogram can be vectorized based on an xy plane in the three-dimensional space. Here, the yz plane in the three-dimensional space can be related to a longitudinal section of the human body, and the xy plane in the three-dimensional space can be related to a transverse section of the human body.

[0037] Next, the data management unit 210 according to an embodiment of the present invention may calculate information regarding a difference between the vectorized data regarding other leads and the vectorized data regarding the specific lead. Specifically, the data management unit 210 according to an embodiment of the present invention may calculate at least one of information regarding a phase difference and information regarding a magnitude difference as information regarding a difference between the vectorized data regarding other leads and the vectorized data regarding the specific lead.

[0038] For example, referring to FIG. 4, the data management unit 210 according to an embodiment of the present invention may vectorize the data on the lead aVR and the data on the leadI as data on other leads and the data on the leadII as data on a specific lead based on the yz plane. The data management unit 210 according to an embodiment of the present invention may calculate θ as information on the phase difference between the vectorized data on the lead aVR and the vectorized data on the leadII, and may calculate θ′ as information on the phase difference between the vectorized data on the leadI and the vectorized data on the leadII. Meanwhile, although not shown in FIG. 4, the data management unit 210 according to an embodiment of the present invention may calculate information on the magnitude difference between the vectorized data on the lead aVR (or the vectorized data on the leadI) and the vectorized data on the leadII when there is a magnitude difference between the vectorized data on the lead aVR (or the vectorized data on the leadI) and the vectorized data on the leadII.

[0039] Next, the data management unit 210 according to one embodiment of the present invention can convert data regarding other leads into enhanced data related to a specific lead by correcting data regarding other leads based on data regarding a specific lead by referring to information regarding the calculated difference.

[0040] Specifically, the data management unit 210 according to an embodiment of the present invention may convert the data of the other lead into augmented data related to the specific lead by correcting the data of the other lead based on the data of the specific lead when the difference between the vectorized data of the other lead and the vectorized data of the specific lead is equal to or greater than a predetermined level. More specifically, the data management unit 210 according to an embodiment of the present invention may convert the data of the other lead into augmented data related to the specific lead by correcting the phase of the data of the other lead based on the phase of the data of the specific lead when the difference between the magnitude of the vectorized data of the other lead and the vectorized data of the specific lead is equal to or greater than a predetermined level. Also, the data management unit 210 according to an embodiment of the present invention may convert the data of the other lead into augmented data related to the specific lead by correcting the magnitude of the data of the other lead based on the magnitude of the data of the specific lead when the difference between the magnitude of the vectorized data of the other lead and the vectorized data of the specific lead is equal to or greater than a predetermined level.

[0041] For example, referring to FIG. 5, the data management unit 210 according to an embodiment of the present invention may determine that the phase difference between the vectorized data related to lead aVR and the vectorized data related to leadII is equal to or greater than a predetermined level (90°), and may convert the data related to lead aVR into augmented data related to the data related to leadII by correcting the phase of the data related to lead aVR based on the phase of the data related to leadII (correcting the phase of the data related to lead aVR by 180°). Here, the augmented data related to the data related to leadII is data related to lead-aVR, and the phase difference (θ″) between the data related to leadII and the data related to lead-aVR may be less than a predetermined level (90°).

[0042] Meanwhile, the data management unit 210 according to an embodiment of the present invention may not correct the data on the other leads if the difference between the vectorized data on the other leads and the vectorized data on the specific lead is less than a predetermined level. In this case, the augmented data related to the specific lead may include the uncorrected data on the other leads.

[0043] For example, referring again to FIG. 4, the data management unit 210 according to an embodiment of the present invention may determine that the phase difference (θ′) between the vectorized data regarding leadI and the vectorized data regarding leadII is less than a predetermined level (90°), and may not correct the phase of the data regarding leadI based on the phase of the data regarding leadII. Here, the augmented data related to the data regarding leadII may be the uncorrected data regarding leadI itself.

[0044] However, in one embodiment of the present invention, the manner in which the data management unit 210 converts data related to other leads into augmented data related to a specific lead is not necessarily limited to the above example, and may be modified in various ways within the scope that can achieve the objectives of the present invention.

[0045] Next, the learning management unit 220 according to one embodiment of the present invention can perform a function of learning an analytical model for discriminating arrhythmia using data related to a specific lead by using augmentation data related to a specific lead and data related to the specific lead as learning data.

[0046] Specifically, according to an embodiment of the present invention, the analysis model trained by the learning management unit 220 can distinguish arrhythmia of the subject by analyzing data related to a specific lead measured from the subject. The learning management unit 220 according to an embodiment of the present invention can train the analysis model using not only data related to a specific lead but also augmented data related to the specific lead as training data. That is, according to an embodiment of the present invention, the analysis model can be trained based on data related to multiple leads (i.e., data related to a specific lead and augmented data related to the specific lead), so that various types of arrhythmia can be distinguished with high accuracy even in an environment where only data related to a single lead (i.e., data related to a specific lead) is input as test input data.

[0047] Next, the communication unit 230 according to an embodiment of the present invention can perform a function of enabling data transmission / reception to / from the data management unit 210 and the learning management unit 220.

[0048] Finally, the control unit 240 according to an embodiment of the present invention can perform a function of controlling data flow between the data management unit 210, the learning management unit 220, and the communication unit 230. That is, the control unit 240 according to the present invention can control the data management unit 210, the learning management unit 220, and the communication unit 230 to perform their respective unique functions by controlling the data flow from / to the outside of the learning data management system 200 or the data flow between each component of the learning data management system 200.

[0049] The above-described embodiments of the present invention may be embodied in the form of program instructions that can be executed by various computer components and recorded on a computer-readable recording medium. The computer-readable recording medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions recorded on the computer-readable recording medium may be specially designed and constructed for the present invention, or may be available to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; and hardware devices specially configured to store and execute program instructions, such as ROMs, RAMs, flash memories, and the like. Examples of program instructions include not only machine code, such as produced by a compiler, but also high-level language code that can be executed by a computer using an interpreter, and the like. The hardware devices may be replaced by one or more software modules to perform the processes of the present invention, and vice versa.

[0050] Although the present invention has been described above using specific details such as specific components and limited examples and drawings, this is merely provided to assist in a more general understanding of the present invention, and the present invention is not limited to the above examples. Those skilled in the art to which the present invention pertains may attempt various modifications and changes from such descriptions.

[0051] Therefore, the concept of the present invention should not be limited to the above described embodiments, and all scopes equivalent to or modified equivalently to the scope of the claims below as well as the claims themselves are within the scope of the concept of the present invention. [Explanation of symbols]

[0052] 100:Communication network 200: Learning Data Management System 210: Data Management Department 220: Learning Management Department 230: Communications Department 240: Control unit 300:Device

Claims

1. 1. A method for managing training data for a biosignal analysis model, comprising: converting data relating to at least one of a plurality of leads into enhanced data relating to a particular lead of the plurality of leads; using the augmentation data and the data related to the particular lead as training data to train an analytical model for discriminating arrhythmia using the data related to the particular lead; The method, wherein the enhanced data includes data obtained by correcting the phase of data for the at least one lead based on the phase of data for the particular lead, or data obtained by correcting the magnitude of data for the at least one lead based on the magnitude of data for the particular lead.

2. The method of claim 1 , wherein the converting step vectorizes the data for the at least one lead and the data for the particular lead.

3. The method of claim 2, wherein the converting step calculates information regarding a difference between the vectorized data for at least one read and the vectorized data for a particular read.

4. The method of claim 3, wherein in the conversion step, data regarding the at least one lead is converted into the enhanced data by correcting data regarding the at least one lead based on data regarding the specific lead by referring to information regarding the calculated difference.

5. The method of claim 4 , wherein the calculated difference information includes at least one of phase difference information and magnitude difference information.

6. A non-transitory computer-readable recording medium having a computer program recorded thereon for carrying out the method according to claim 1.

7. A system for managing learning data for a biological signal analysis model, comprising: a data manager that converts data relating to at least one of a plurality of leads into augmented data relating to a particular lead of the plurality of leads; and a learning management unit that uses the enhancement data and the data related to the specific lead as learning data to train an analysis model for discriminating arrhythmia using the data related to the specific lead; The system, wherein the enhanced data includes data obtained by correcting the phase of data for the at least one lead based on the phase of data for the particular lead, or includes data obtained by correcting the magnitude of data for the at least one lead based on the magnitude of data for the particular lead.

8. The system of claim 7 , wherein the data manager vectorizes the data regarding the at least one lead and the data regarding the particular lead.

9. The system of claim 8 , wherein the data management unit calculates information regarding a difference between the vectorized data for at least one lead and the vectorized data for the particular lead.

10. The system of claim 9 , wherein the data management unit converts the data regarding the at least one lead into the enhanced data by correcting the data regarding the at least one lead based on the data regarding the specific lead, by referring to information regarding the calculated difference.

11. The system of claim 10 , wherein the calculated difference information includes at least one of phase difference information and magnitude difference information.

Images