Apparatus, method, and computer program for predicting left ventricular diastolic dysfunction of user on basis of electrocardiogram data

Abstract

The present disclosure provides an apparatus, a method, and a computer program for predicting left ventricular diastolic dysfunction on the basis of electrocardiogram data. The method according to one embodiment of the present disclosure comprises the steps of: evaluating a left ventricular diastolic function of a subject; and obtaining training data by performing labeling on electrocardiogram data of the subject on the basis of an evaluation result of the left ventricular diastolic function of the subject.

Description

Device, method and computer program for predicting left ventricular diastolic dysfunction of a user based on electrocardiogram data

[0001] The present disclosure relates to deep learning technology in the medical field, and more particularly, to a device, method, and computer program for predicting left ventricular diastolic dysfunction of a user based on electrocardiogram data using deep learning technology.

[0002] With recent advancements in information and communication technology and artificial intelligence algorithms, the medical field is actively researching technologies that analyze or predict diseases based on a patient's biosignals. Among these, electrocardiograms (ECGs) are widely used for the diagnosis and monitoring of heart disease, as they can noninvasively measure the electrophysiological activity of the heart. In particular, ECGs are gaining attention as an early diagnosis tool for various heart diseases due to their ease of measurement, low cost, and ability to reflect changes in cardiac function over time.

[0003] The relaxation function of the left ventricle, a cardiac function, is essential for the heart to effectively receive blood. When this function declines, it leads to left ventricular diastolic dysfunction (LVDD). LVDD is caused by a variety of factors, including hypertension, diabetes, and aging. Initially, the absence of clear symptoms makes diagnosis difficult. However, if left untreated, it can progress to heart failure, with symptoms such as pulmonary congestion and dyspnea. Therefore, early detection and management are crucial. Therefore, there is a growing need for technological approaches that can assess and predict LVDD early, based on noninvasive biosignals such as electrocardiograms.

[0004] The present disclosure, conceived in response to the aforementioned background technology, aims to provide a device, method, and computer program for predicting a user's left ventricular diastolic dysfunction based on electrocardiogram data. However, the problems addressed by the present disclosure are not limited to the problems mentioned above, and other problems not mentioned will be clearly understood based on the description below.

[0005] A method for predicting left ventricular diastolic dysfunction based on electrocardiogram data, which is performed by a computing device including at least one processor for solving a task as described above, includes a step of obtaining an evaluation result regarding the left ventricular diastolic function of a subject, and a step of labeling the electrocardiogram data of the subject based on the evaluation result regarding the left ventricular diastolic function of the subject to obtain learning data.

[0006] Alternatively, the evaluating step comprises evaluating the left ventricular diastolic function of the subject based on the ratio of the early diastolic mitral annulus blood flow velocity (E) to the early diastolic mitral annulus tissue velocity (e') of the subject.

[0007] Alternatively, the step of acquiring the learning data includes a step of acquiring label data included in the learning data by assigning a label value corresponding to the presence of the left ventricular diastolic dysfunction to the electrocardiogram data if the ratio is greater than or equal to a first reference value, and by assigning a label value corresponding to the absence of the left ventricular diastolic dysfunction to the electrocardiogram data if the ratio is less than or equal to the first reference value.

[0008] Alternatively, the method comprises a step of training a first neural network model to produce a score corresponding to the likelihood of left ventricular diastolic dysfunction of the subject based on the learning data.

[0009] Alternatively, the step of assessing left ventricular diastolic function of the subject based on the ratio comprises: calculating a first ratio of early diastolic mitral annulus blood flow velocity (E) to early diastolic mitral annulus tissue velocity (e') based on early diastolic mitral annulus tissue velocity (e') measured at the interventricular septum of the subject; calculating a second ratio of early diastolic mitral annulus blood flow velocity (E) to early diastolic mitral annulus tissue velocity (e') based on early diastolic mitral annulus tissue velocity (e') measured at the lateral heart wall of the subject; and assessing left ventricular diastolic function of the subject based on at least one of the first and second ratios.

[0010] Alternatively, the step of obtaining the learning data includes a step of obtaining label data included in the learning data by assigning the ratio as a label value corresponding to the electrocardiogram data.

[0011] Alternatively, the method comprises a step of training a second neural network model to calculate a ratio of the early diastolic mitral annulus blood flow velocity (E) to the early diastolic mitral annulus tissue velocity (e') of the subject based on the learning data.

[0012] A computing device for predicting left ventricular diastolic dysfunction based on electrocardiogram data for solving the aforementioned problem comprises a processor including at least one core and a memory including program codes executable by the processor, wherein the processor obtains an evaluation result regarding the left ventricular diastolic function of a subject and performs labeling on the electrocardiogram data of the subject based on the evaluation result regarding the left ventricular diastolic function of the subject to obtain learning data.

[0013] A computer program stored in a computer-readable storage medium for solving the aforementioned problem, wherein the computer program, when executed on one or more processors, performs an operation for predicting left ventricular diastolic dysfunction based on electrocardiogram data, the operation including a step of obtaining an evaluation result regarding the left ventricular diastolic function of a subject and a step of labeling the electrocardiogram data of the subject based on the evaluation result regarding the left ventricular diastolic function of the subject to obtain learning data.

[0014] According to one embodiment of the present disclosure, left ventricular diastolic dysfunction can be predicted in advance by labeling electrocardiogram data based on assessment results of left ventricular diastolic function and using this data to train a neural network model. This allows for early detection of left ventricular diastolic dysfunction without the need for separate, expensive equipment or complex image analysis, thereby preventing progression to serious heart diseases such as heart failure.

[0015] FIG. 1 is an exemplary diagram of a computing device that predicts left ventricular diastolic dysfunction of a user based on electrocardiogram data according to one embodiment of the present disclosure.

[0016] FIG. 2 is a block diagram of a computing device according to an embodiment of the present disclosure.

[0017] FIG. 3 is a flowchart of a method for performing a clinical trial based on electrocardiogram data according to an embodiment of the present disclosure.

[0018] FIG. 4 is an exemplary diagram of a method for preparing label data as learning data based on the ratio of early diastolic mitral annulus blood flow velocity (E) to early diastolic mitral annulus tissue velocity (e') according to one embodiment of the present disclosure.

[0019] FIG. 5 is a flowchart of a method for determining a user's status based on a pre-learned neural network model according to an embodiment of the present disclosure.

[0020] FIG. 6 is an exemplary diagram of a method for predicting left ventricular diastolic dysfunction of a user based on a first neural network model and a second neural network model according to an embodiment of the present disclosure.

[0021] FIG. 7 is a block diagram of a computing device according to another embodiment of the present disclosure.

[0022] Below, embodiments of the present disclosure are described in detail with reference to the attached drawings so that those skilled in the art can easily implement the present disclosure. The embodiments presented in this disclosure are provided to enable those skilled in the art to utilize or implement the contents of the present 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 may be implemented in various different forms and is not limited to the embodiments described below.

[0023] Throughout the specification of this disclosure, identical or similar drawing numbers refer to identical or similar components. Furthermore, for clarity in the description of this disclosure, drawing numbers for parts unrelated to the description of this disclosure may be omitted in the drawings.

[0024] The term "or" as used herein is intended to mean an inclusive "or" rather than an exclusive "or." That is, unless otherwise specified herein or clear from context, "X employs A or B" should be understood to mean either of its natural inclusive permutations. For example, unless otherwise specified herein or clear from context, "X employs A or B" can be interpreted to mean either X employs A, X employs B, or X employs both A and B.

[0025] The term "and / or" as used herein should be understood to refer to and include all possible combinations of one or more of the related concepts listed.

[0026] The terms "comprises" and / or "comprising" as used herein should be understood to mean the presence of certain features and / or components. However, it should be understood that the terms "comprises" and / or "comprising" do not exclude the presence or addition of one or more other features, other components, and / or combinations thereof.

[0027] Unless otherwise specified in this disclosure or unless the context makes it clear that the singular form is being referred to, the singular should generally be construed to include “one or more.”

[0028] The term "Nth (N is a natural number)" used in this disclosure can be understood as an expression used to mutually distinguish components of this disclosure based on a predetermined standard such as a functional perspective, a structural perspective, or convenience of explanation. For example, components performing different functional roles in this disclosure can be distinguished as a first component or a second component. However, components that are substantially the same within the technical spirit of this disclosure but must be distinguished for convenience of explanation may also be distinguished as a first component or a second component.

[0029] The term "acquisition" as used in this disclosure may be 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.

[0030] Meanwhile, the term "module" or "unit" used in the present disclosure can be understood as a term referring to an independent functional unit that processes computing resources, such as a computer-related entity, firmware, software or a part thereof, hardware or a part thereof, or a combination of software and hardware. At this time, the "module" or "unit" may be a unit composed of a single element, or a unit expressed as a combination or set of multiple elements. For example, as a narrow concept, a "module" or "unit" may refer to a hardware element of a computing device or a set thereof, an application program that performs a specific function of software, a processing process implemented through software execution, or a set of instructions for program execution, etc. In addition, as a broad concept, a "module" or "unit" may refer to the computing device itself that constitutes the system, or an application running on the computing device, etc. However, since the above-described concept is only an example, the concept of “module” or “part” may be defined in various ways within a range understandable to those skilled in the art based on the contents of the present disclosure.

[0031] The term "model" as used herein may be understood as a system implemented using mathematical concepts and language to solve a specific problem, a set of software units to solve a specific problem, or an abstract model of a processing process to solve a specific problem. For example, a neural network "model" may refer to the entire system implemented as a neural network that has problem-solving capabilities through learning. In this case, the neural network can have problem-solving capabilities by optimizing the parameters connecting nodes or neurons through learning. A neural network "model" may include a single neural network or a set of neural networks that are a combination of multiple neural networks.

[0032] The term "data" used in this disclosure may include "images," signals, and the like. The term "image" used in this disclosure may refer to multidimensional data composed of discrete image elements. In other words, "image" may be understood as a term referring to a digital representation of an object visible to the human eye. For example, "image" may refer to multidimensional data composed of elements corresponding to pixels in a two-dimensional image. "Image" may refer to multidimensional data composed of elements corresponding to voxels in a three-dimensional image.

[0033] The explanation of the above terms is intended to aid understanding of the present disclosure. Therefore, unless explicitly stated as limiting the contents of the present disclosure, it should be noted that the above terms are not intended to limit the technical ideas of the contents of the present disclosure.

[0034] FIG. 1 is an exemplary diagram of a computing device that predicts left ventricular diastolic dysfunction of a user based on electrocardiogram data (20) according to one embodiment of the present disclosure.

[0035] A computing device (100) according to an embodiment of the present disclosure may be a hardware device or a 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 interface. For example, the computing device (100) may be a server that performs intensive data processing functions and shares resources, or may be a client that shares resources through interaction with a server. In addition, 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 one example related to the type of the computing device (100), the type of the computing device (100) may be configured in various ways within a range that can be understood by those skilled in the art based on the contents of the present disclosure.

[0036] Referring to FIG. 1, a computing device (100) obtains electrocardiogram data (20) from a user (1), and determines the possibility of left ventricular diastolic dysfunction of the user (1) based on the electrocardiogram data (20) of the user (1) using a pre-learned neural network model (600). Here, the possibility of left ventricular diastolic dysfunction includes whether left ventricular diastolic dysfunction is currently occurring or is likely to occur in the future. In particular, the computing device (100) can input the electrocardiogram data (20) of the user (1) into the pre-learned neural network model (600) to obtain a score corresponding to the possibility of left ventricular diastolic dysfunction. The score may be a value quantitatively representing the risk of left ventricular diastolic dysfunction of the user (1) inferred by the neural network model (600) based on feature information extracted from the electrocardiogram data (20). The computing device (100) may obtain a score from a pre-trained neural network model (600), or may obtain a score by applying a weight (e.g., a weight set according to the biological characteristics of the subject (2)) or a pre-set formula to output data obtained from the pre-trained neural network model (600) (e.g., a probability value regarding left ventricular diastolic dysfunction calculated from a classifier of the neural network model (600)). The higher the score, the more likely the computing device (100) is to determine that the user (1) has left ventricular diastolic dysfunction. The computing device (100) may determine whether there is an abnormality in the heart of the user (1) based on the obtained score, or may monitor changes in the health status of the user (1) by comparing it with scores obtained in the past.

[0037] Meanwhile, in order to obtain a score corresponding to the possibility of left ventricular diastolic dysfunction through a neural network model (600), the computing device (100) may train the neural network model (600) in advance based on learning data composed of a plurality of electrocardiogram data (20) obtained from a plurality of subjects (2). At this time, the learning data may include label data in which a label date related to left ventricular diastolic dysfunction is assigned to a plurality of electrocardiogram data (20). The computing device (100) may evaluate the cardiac function of each subject (2) that measured the electrocardiogram data (20) and assign a label to the electrocardiogram data (20) based on the evaluation result to prepare label data.

[0038] In particular, the computing device (100) can calculate E / e', which is the ratio of the early diastolic mitral annulus blood flow velocity (E) to the early diastolic mitral annulus tissue velocity (e'), and label the electrocardiogram data (20) based on the ratio. E / e' is an index that can be directly calculated through an echocardiographic examination, and has the advantage of being able to perform labeling according to objective criteria, excluding the subjective judgment of medical staff. Accordingly, the computing device (100) can generate label data based on a consistent standard, and this labeling method can improve the reliability and consistency of learning data. Hereinafter, an embodiment of the present disclosure related thereto will be described in detail with reference to FIGS. 2 to 6. Meanwhile, the contents of the present disclosure described above with respect to FIG. 1 can be equally applied to the following embodiments.

[0039] FIG. 2 is a block diagram of a computing device (100) according to one embodiment of the present disclosure.

[0040] Referring to FIG. 2, a computing device (100) according to an embodiment of the present disclosure may include a processor (110) and a memory (120). However, FIG. 2 is merely an example, and the computing device (100) may include other components for implementing a computing environment. In addition, only some of the components disclosed in FIG. 1 may be included in the computing device (100).

[0041] A processor (110) according to an embodiment of the present disclosure may be understood as a configuration unit including hardware and / or software for performing computing operations. For example, the processor (110) may read a computer program to perform data processing for machine learning. The processor (110) may process computational processes such as processing input data for machine learning, feature extraction for machine learning, and error calculation based on backpropagation. The processor (110) for performing such data processing may include a central processing unit (CPU), a general purpose graphics processing unit (GPGPU), a tensor processing unit (TPU), an application specific integrated circuit (ASIC), or a field programmable gate array (FPGA). The above-described type of processor (110) is only one example, and thus, 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.

[0042] The processor (110) is connected to other components of the computing device (100), i.e., the memory (120), and controls the overall operation of the computing device (100).

[0043] The memory (120) according to one embodiment of the present disclosure may be understood as a configuration unit including hardware and / or software for storing and managing data processed in the computing device (100). That is, the memory (120) may store any type of data generated or determined by the processor (110) and any type of data received by the communication interface (130). For example, the memory (120) may include at least one type of storage medium among a flash memory type, a hard disk type, a multimedia card micro type, a card type memory, a random access memory (RAM), a static random access memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, and an optical disk. In addition, the memory (120) may also include a database system that controls and manages data in a predetermined system. The type of memory (120) described above is only one example, and thus the type of memory (120) can be configured in various ways within a range understandable to those skilled in the art based on the contents of the present disclosure.

[0044] The memory (120) can structure and organize and manage data, combinations of data, and program codes executable in the processor (110) required for the processor (110) to perform operations. For example, the memory (120) can store a neural network model (600) trained to predict left ventricular diastolic dysfunction of a user (1) based on electrocardiogram data (20), and can store program codes that operate the neural network model (600) to perform learning, program codes that operate the neural network model (600) to receive electrocardiogram data (20) and perform inference in accordance with the purpose of use of the computing device (100) (for example, for the purpose of calculating a score for a specific disease), and output or processed data generated as the program code is executed. In addition, the memory (120) can also store learning data for training the neural network model (600).

[0045] FIG. 3 is a flowchart of a method for performing a clinical trial based on electrocardiogram data (20) according to one embodiment of the present disclosure.

[0046] According to one embodiment of the present disclosure, the processor (110) can obtain an evaluation result regarding the left ventricular diastolic function of the subject (2) (S310). To this end, the processor (110) can evaluate the left ventricular diastolic function of the subject (2). Specifically, the processor (110) can evaluate the left ventricular diastolic function of the subject (2) to determine a label value or a type of label assigned to the electrocardiogram data (20) prior to acquiring learning data. Here, the subject (2) may be the same subject (2) as the measurement target of the acquired electrocardiogram data (20). When the electrocardiogram data (20) of the subject (2) is acquired, the processor (110) can evaluate the left ventricular diastolic function of the subject (2). However, this is not limited thereto, and the acquisition and evaluation order may be set in various ways.

[0047] The electrocardiogram data (20) may include at least one of 12-lead electrocardiogram data (20), 6-lead electrocardiogram data (20), and single-lead electrocardiogram data (20). The processor (110) may configure the acquired electrocardiogram data (20) as input data of learning data of the neural network model (600) after undergoing preprocessing processes such as normalization, noise removal, and segmentation through a digital signal processing process. For example, the processor (110) may obtain the electrocardiogram data (20) by sampling the electrocardiogram signal at 500 points per second (Sampling rate = 500 Hz) and dividing it into segments of 8 seconds in length.

[0048] At this time, the processor (110) can evaluate the left ventricular relaxation function of the subject (2) by using a sensing unit (e.g., an ultrasonic probe) included in the computing device (100). Alternatively, the left ventricular relaxation function of the subject (2) can be evaluated by obtaining data on the relaxation function of the subject (2) measured by an external biosignal measuring device (e.g., an ultrasonic device) through a communication interface of the computing device (100).

[0049] The processor (110) can directly obtain electrocardiogram data (20) of the subject (2) using a sensing unit (e.g., multiple electrodes) of the computing device (100), or can obtain electrocardiogram data (20) of the subject (2) measured by an external biosignal measuring device (e.g., an electrocardiogram device) through a communication interface of the computing device (100).

[0050] Assessing left ventricular diastolic function can be done by quantitatively or qualitatively assessing the blood-holding capacity of the left ventricle during the post-systolic relaxation phase of the heart to determine whether left ventricular diastolic function is impaired or whether left ventricular diastolic dysfunction exists.

[0051] According to one embodiment of the present disclosure, the processor (110) can evaluate the left ventricular diastolic function of the subject (2) based on the ratio of the early diastolic mitral annular velocity (e' velocity, hereinafter, e') to the early diastolic mitral inflow velocity (E velocity, hereinafter, E) of the subject (2). The ratio of the early diastolic mitral annular velocity (e') to the early diastolic mitral annular inflow velocity (E) of the subject (2) can be calculated based on the early diastolic mitral annular blood flow velocity (E) and the early diastolic mitral annular tissue velocity (e') values ​​measured through echocardiography.

[0052] The processor (110) may determine that the subject (2) has left ventricular diastolic dysfunction if the ratio is greater than or equal to a first reference value. Conversely, the processor (110) may determine that the subject (2) does not have left ventricular diastolic dysfunction if the ratio is less than or equal to the first reference value. For example, the processor (110) may determine that the subject (2) has left ventricular diastolic dysfunction if the ratio is greater than or equal to 15, and may determine that the subject (2) does not have left ventricular diastolic dysfunction if the ratio is less than 15.

[0053] Meanwhile, the first reference value may be set differently based on the age, sex, and presence or absence of an underlying disease of the subject (2), and the reference value for the presence of left ventricular diastolic function and the reference value for the absence of left ventricular diastolic function may also be set differently.

[0054] In calculating the ratio of the early diastolic mitral annulus blood flow velocity (E) to the early diastolic mitral annulus tissue velocity (e') according to one embodiment of the present disclosure, the processor (110) may calculate a first ratio of the early diastolic mitral annulus blood flow velocity (E) to the early diastolic mitral annulus tissue velocity (e') based on the early diastolic mitral annulus tissue velocity (e') measured at the interventricular septum of the subject (2), and may calculate a second ratio of the early diastolic mitral annulus blood flow velocity (E) to the early diastolic mitral annulus tissue velocity (e') measured at the lateral heart wall of the subject (2). In addition, the processor (110) may evaluate the left ventricular diastolic function of the subject (2) based on at least one of the first and second ratios.

[0055] Specifically, the processor (110) can calculate a first ratio (E / e' septal) using the early diastolic mitral annulus tissue velocity (e' septal) and the early diastolic mitral annulus blood flow velocity (E) measured at the septal wall of the subject (2). At this time, E is the early inflow velocity of the left ventricular filling flow through the mitral valve measured by Doppler echocardiography. In addition, the processor (110) can calculate a second ratio (E / e' lateral) using the same E value as the early diastolic mitral annulus tissue velocity (e' lateral) measured at the lateral wall of the heart of the subject (2). In addition, the processor (110) can evaluate the left ventricular diastolic function by selectively utilizing one of the first ratio and the second ratio, or by calculating an average value of the two ratios. For example, the processor (110) may determine that the subject (2) has left ventricular diastolic dysfunction if the average ratio is greater than or equal to a first reference value, and may determine that the subject (2) does not have left ventricular diastolic dysfunction if the average ratio is less than or equal to a second reference value. In this way, the processor (110) may obtain more reliable evaluation results regarding left ventricular diastolic function by reflecting the velocities measured at different parts of the heart, thereby compensating for measurement errors or deviations at individual parts.

[0056] The processor (110) can obtain learning data by labeling the electrocardiogram data (20) of the subject (2) based on the evaluation results regarding the left ventricular diastolic function of the subject (2) (S320). Specifically, the processor (110) can obtain label data that is assigned a label regarding the presence or absence of left ventricular diastolic dysfunction based on the evaluation results.

[0057] FIG. 4 is an exemplary diagram of a method for preparing label data as learning data based on the ratio of early diastolic mitral annulus blood flow velocity (E) to early diastolic mitral annulus tissue velocity (e') according to one embodiment of the present disclosure.

[0058] In particular, referring to FIG. 4, the processor (110) may assign a label value (e.g., 1) corresponding to the presence of left ventricular diastolic dysfunction to the electrocardiogram data (20) if the ratio of the early diastolic mitral annulus blood flow velocity (E) to the early diastolic mitral annulus tissue velocity (e') is greater than or equal to a first reference value (e.g., 15). In addition, the processor (110) may assign a label value (e.g., 0) corresponding to the absence of left ventricular diastolic dysfunction to the electrocardiogram data (20) if the ratio of the early diastolic mitral annulus blood flow velocity (E) to the early diastolic mitral annulus tissue velocity (e') is less than the first reference value.

[0059] FIG. 5 is a flowchart of a method for identifying a user's status based on a pre-trained neural network model according to an embodiment of the present disclosure. S510 and S520 illustrated in FIG. 5 may correspond to S310 and S320 illustrated in FIG. 3 , respectively.

[0060] In addition, the processor (110) can train a neural network model (600) based on the learning data (60) (S530). According to one embodiment of the present disclosure, the neural network model (600) may include a first neural network model (610) trained to produce a score corresponding to the likelihood of left ventricular diastolic dysfunction of the subject (2).

[0061] The processor (110) inputs a plurality of electrocardiogram data (20) included in the learning data (60) as input data to a neural network model (600), and can calculate a loss function based on the difference between the output value of each electrocardiogram data (20) of the neural network model (600) and the label assigned to the label data (e.g., presence or absence of left ventricular diastolic dysfunction). The loss function can be defined as a cross entropy loss or an objective function that optimizes the balance between precision and recall. Based on the calculated loss function, the processor (110) can adjust the weights of the model (600) through backpropagation. By repeating this process, the processor (110) can improve the classification performance of the neural network model (600) regarding left ventricular diastolic dysfunction of the user (1), and finally obtain a first neural network model (610) trained to produce a score corresponding to the possibility of left ventricular diastolic dysfunction of the user (1) based on electrocardiogram data (20).

[0062] The neural network model (600) can be implemented in a multi-layer perceptron (MLP), convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks, or residual neural networks (ResNet) structure.

[0063] According to one embodiment of the present disclosure, the first neural network model (610) may include multiple neural network models corresponding to the acquisition method of electrocardiogram data (e.g., 12-lead method, 6-lead method, single-lead method, etc.). To this end, the processor (110) may prepare training data by dividing the plurality of electrocardiogram data acquired according to each acquisition method and train each neural network model.

[0064] Meanwhile, when the learning of the neural network model (600) is completed, the processor (110) can identify the status of the user (1) using the learned neural network model (600) (S540).

[0065] In particular, the processor (110) may analyze the electrocardiogram data (20) using the first neural network model (610) and obtain a score corresponding to the possibility of left ventricular diastolic dysfunction. At this time, if the score is equal to or greater than a preset score, the processor (110) may determine that the user (1) has left ventricular diastolic dysfunction or that the user (1) has a high possibility of developing left ventricular diastolic dysfunction. Accordingly, the processor (110) may provide information guiding appropriate action. For example, the processor (110) may recommend a detailed echocardiographic examination to the user (1) or output a message inducing a medical consultation with a relevant specialist based on the determination result. Meanwhile, if the score is less than a preset score, the processor (110) may determine that the user (1) does not have left ventricular diastolic dysfunction or that the user (1) has a low possibility of developing left ventricular diastolic dysfunction.

[0066] In addition, the processor (110) can monitor the change trend of the score for the same user (1) to quantitatively evaluate the progression of left ventricular diastolic dysfunction or analyze the trend. For example, if the score continuously increases over a certain period of time, the processor (110) can determine that there is a high possibility that the heart function of the user (1) is gradually deteriorating, and accordingly, can provide a warning message to the medical staff or the user (1) notifying them of the need for periodic monitoring and early intervention.

[0067] Meanwhile, according to one embodiment of the present disclosure, the processor (110) may obtain label data included in the learning data (60) by assigning the ratio of the early diastolic mitral annulus blood flow velocity (E) to the early diastolic mitral annulus tissue velocity (e') as a label value corresponding to the electrocardiogram data (20).

[0068] Specifically, the processor (110) may directly assign a label value to the electrocardiogram data (20) as a continuous value, which is the ratio of the early diastolic mitral annulus blood flow velocity (E) to the early diastolic mitral annulus tissue velocity (e'). That is, the label may include an E / e' value, which is a quantitative index, in the form of a regression. Accordingly, the label data may be composed of actual E / e' values ​​corresponding to each electrocardiogram data (20). In addition, the processor (110) may train a neural network model (600) based on the learning data (60). According to one embodiment of the present disclosure, the neural network model (600) may include a second neural network model (620) trained to calculate the ratio of the early diastolic mitral annulus blood flow velocity (E) to the early diastolic mitral annulus tissue velocity (e') of the subject (2).

[0069] The processor (110) may provide a plurality of electrocardiogram data (20) included in the learning data (60) as input data to the second neural network model (620), and may calculate a loss function based on the difference between the E / e' value for each electrocardiogram data (20) output from the second neural network model (620) and the assigned label (e.g., the E / e' value). The loss function may be defined as a loss function suitable for a regression problem, such as a mean squared error (MSE) or a mean absolute error (MAE). Based on the loss function, the processor (110) may adjust the weights of the second neural network model (620) through backpropagation. By repeating this process, the processor (110) can improve the performance of calculating the E / e' value of the second neural network model (620), and ultimately obtain a regression-based second neural network model (620) trained to calculate the E / e' value based on electrocardiogram data (20).

[0070] According to one embodiment of the present disclosure, the second neural network model (620) may include multiple neural network models corresponding to the acquisition method of electrocardiogram data (e.g., 12-lead method, 6-lead method, single-lead method, etc.). In this regard, the description of the first neural network model (610) described above may be equally applicable.

[0071] Meanwhile, when the learning of the neural network model (600) is completed, the processor (110) may analyze the electrocardiogram data (20) of the user (1) using the previously learned neural network model (600) (i.e., the second neural network model (620)) and obtain a ratio of the early diastolic mitral annulus blood flow velocity (E) to the early diastolic mitral annulus tissue velocity (e'). At this time, if the obtained ratio is 15 or more, the processor (110) may determine that the user (1) has left ventricular diastolic dysfunction or that the user (1) is likely to develop left ventricular diastolic dysfunction. On the other hand, if the obtained ratio is less than 15, the processor (110) may determine that the user (1) does not have left ventricular diastolic dysfunction or that the user (1) is unlikely to develop left ventricular diastolic dysfunction.

[0072] Meanwhile, according to one embodiment of the present disclosure, the neural network model (600) (i.e., the first and second neural network models (610 and 620)) may include a plurality of sub-neural network models depending on the type of electrocardiogram data (20). Specifically, the neural network model (600) may include sub-neural network models each corresponding to a 12-lead electrocardiogram, a 6-lead electrocardiogram, or a single-lead electrocardiogram depending on the lead method in which the input electrocardiogram data (20) is acquired. The processor (110) may identify the lead type of the input electrocardiogram data (20) and select a sub-neural network model trained to be optimized for the corresponding lead method to input the electrocardiogram data (20). At this time, the training data (60) of each sub-neural network model may be configured in a form in which a plurality of electrocardiogram data (20) included in the electrocardiogram data (20) are combined depending on the lead method.

[0073] Meanwhile, the processor (110) can input the biological information (age, weight, gender, etc.) of the user (1) who acquired the electrocardiogram data (20) together with the electrocardiogram data (20) into the neural network model (600) (i.e., the first and second neural network models (610 and 620)) to obtain a score or ratio. To this end, the neural network model (600) (the first and second neural network models (610 and 620)) can be trained with the biological information of the subject (2) together with the electrocardiogram data as auxiliary input.

[0074] FIG. 6 is an exemplary diagram of a method for predicting left ventricular diastolic dysfunction of a user (1) based on a first neural network model (610) and a second neural network model (620) according to one embodiment of the present disclosure.

[0075] Meanwhile, according to one embodiment of the present disclosure, the processor (110) can predict left ventricular diastolic dysfunction of the user (1) using the first neural network model (610) and the second neural network model (620). In this regard, referring to FIG. 6, if the score obtained from the first neural network model (610) is equal to or greater than a preset first score, the processor (110) can determine that the user (1) has left ventricular diastolic dysfunction or that the user (1) is likely to develop left ventricular diastolic dysfunction. At this time, if the score obtained from the first neural network model (610) is less than the preset first score and greater than or equal to the second score, the processor (110) can determine the presence or possibility of left ventricular diastolic dysfunction in the user (1) based on the ratio of the early diastolic mitral annulus blood flow velocity (E) to the early diastolic mitral annulus tissue velocity (e') obtained through the second neural network model (620). Meanwhile, if the score obtained from the first neural network model (610) is less than the preset second score, the processor (110) can determine that the user (1) does not have left ventricular diastolic dysfunction or that the user (1) is unlikely to have left ventricular diastolic dysfunction.

[0076] Fig. 7 is a block diagram of a computing device (700) according to another embodiment of the present disclosure. Referring to Fig. 7, a computing device (700) according to an embodiment of the present disclosure includes a processor (710), a memory (720), a communication interface (730), a sensing unit (740), a display (750), a user interface (760), a camera (770), and a speaker (780). Among the configurations illustrated in Fig. 7, the processor (710), the memory (720), and the communication interface (730) correspond to the configurations of the processor (110) and the memory (120) of the computing device (100) illustrated in Fig. 2, and thus a detailed description thereof will be omitted.

[0077] A communication interface (130) according to an embodiment of the present disclosure may be understood as a component that transmits and receives data through any known wired or wireless communication system. For example, the communication interface (130) may perform data transmission and reception using a wired or wireless communication system such as a local area network (LAN), wideband code division multiple access (WCDMA), long term evolution (LTE), wireless broadband internet (WiBro), 5th generation mobile communication (7G), ultrawide-band, ZigBee, radio frequency (RF) communication, wireless LAN, wireless fidelity, near field communication (NFC), or Bluetooth. Since the above-described communication systems are only examples, the wired and wireless communication system for data transmission and reception of the communication interface (130) may be applied in various ways other than the above-described examples.

[0078] The communication interface (130) can receive data necessary for the processor (110) to perform calculations through wired or wireless communication with any system or any client, etc. In addition, the communication interface (130) can transmit data generated through calculations of the processor (110) through wired or wireless communication with any system or any client, etc. For example, the communication interface (130) can receive electrocardiogram data (20) of a user (1) or an evaluation result regarding left ventricular contraction function through communication with a biosignal measuring device. The communication interface (130) can transmit output data of the neural network model (700), intermediate data, processed data, etc. derived from the calculation process of the processor (110), etc. through communication with the aforementioned database, server, or computing device, etc.

[0079] The sensing unit (740) can obtain electrocardiogram data (20) for the user (1). For example, the sensing unit (740) can include a plurality of electrodes. At this time, the processor (110) can obtain the electrocardiogram data (20) of the user (1) as the electrocardiogram data (20) through at least one electrode. Alternatively, the processor (110) can obtain the ratio of the early diastolic mitral annulus blood flow velocity (E) to the early diastolic mitral annulus tissue velocity (e') of the user (1) by using an ultrasound probe included in the sensing unit (740).

[0080] The display (750) can display various images. Here, the images include both still images and moving images. The display (750) can output the result of determining the status of the user (1) or output information related to an electrocardiogram (such as an electrocardiogram graph). The display (750) can be implemented as a display in various forms, such as an LCD (Liquid Crystal Display Panel), an OLED (Organic Light Emitting Diodes), an LCoS (Liquid Crystal on Silicon), a DLP (Digital Light Processing), etc. In addition, the display (750) can also include a driving circuit, a backlight unit, etc., which can be implemented in a form, such as an a-TFT, an LTPS (low temperature poly silicon) TFT, an OTFT (organic TFT), etc.

[0081] Meanwhile, the display (750) may be implemented as a touch screen by being combined with a touch panel, and in this case, the display (750) may perform the function of not only an output interface that outputs an image through the touch screen, but also an input interface that receives a touch input from a user (1).

[0082] The user interface (760) is a component used by the computing device (700) to interact with the user, and may include, but is not limited to, at least one of a touch sensor, a motion sensor, a button, a switch, and a jog dial. The processor (710) may receive input of the user's underlying disease and biological information through the user interface (760).

[0083] The camera (770) captures an object around the computing device (700) to obtain an image of the object. Specifically, the camera (770) can capture an image of the user (1). At this time, the processor (710) can identify the underlying disease of the user (1) based on the captured image. To this end, the camera (770) may be implemented with an imaging device such as a CMOS image sensor (CIS) having a CMOS structure, a charge coupled device (CCD) having a CCD structure, etc. However, the present invention is not limited thereto, and the camera (770) may be implemented with a camera module of various resolutions capable of capturing an object. Meanwhile, the camera (770) may be implemented with a depth camera (e.g., an IR depth camera), a stereo camera, an RGB camera, etc.

[0084] The speaker (780) is a component that outputs various audio data on which various processing operations such as decoding, amplification, and noise filtering have been performed by an audio processing unit (not shown). The speaker (780) can output various notification sounds or voice messages. According to one embodiment of the present disclosure, the processor (710) can convert an electrical signal received from an external device into a user's (1) voice and output it through the speaker (780). For example, the speaker (780) can output a judgment result and guide information regarding the user's (1) status in the form of a voice message.

[0085] The various embodiments of the present disclosure described above can be combined with additional embodiments and modified within the scope understood by those skilled in the art in light of the detailed description above. It should be understood that the embodiments of the present disclosure are illustrative in all respects and not restrictive. For example, each component described as a single component may be implemented in a distributed manner, and likewise, components described as distributed may be implemented in a combined manner. Accordingly, all changes or modifications derived from the meaning, scope, and equivalent concepts of the claims of the present disclosure should be construed as being included within the scope of the present disclosure.

Claims

1. A method for predicting left ventricular diastolic dysfunction based on electrocardiogram data performed by a computing device including at least one processor, A step of obtaining evaluation results regarding the left ventricular diastolic function of the subject; and A step of obtaining learning data by labeling the electrocardiogram data of the subject based on the evaluation result of the left ventricular diastolic function of the subject; including; method.

2. In paragraph 1, The step of obtaining the above evaluation results is: A step of evaluating the left ventricular diastolic function of the subject based on the ratio of the early diastolic mitral annulus blood flow velocity (E velocity) to the early diastolic mitral annulus tissue velocity (e' velocity) of the subject; including; method.

3. In paragraph 2, The step of acquiring the above learning data is: A step of obtaining label data included in the learning data by assigning a label value corresponding to the presence of left ventricular diastolic dysfunction to the electrocardiogram data if the ratio is greater than or equal to the first reference value, and assigning a label value corresponding to the absence of left ventricular diastolic dysfunction to the electrocardiogram data if the ratio is less than or equal to the first reference value; method.

4. In paragraph 3, A step of training a first neural network model to produce a score corresponding to the possibility of left ventricular diastolic dysfunction of the subject based on the above learning data; method.

5. In paragraph 2, The step of evaluating the left ventricular diastolic function of the subject based on the above ratio is: A step of calculating a first ratio of the early diastolic mitral annulus blood flow velocity (E velocity) to the early diastolic mitral annulus tissue velocity (e' velocity) based on the early diastolic mitral annulus tissue velocity (e' velocity) measured in the interventricular septum of the subject; A step of calculating a second ratio of the early diastolic mitral annulus blood flow velocity (E velocity) to the early diastolic mitral annulus tissue velocity (e' velocity) based on the early diastolic mitral annulus tissue velocity (e' velocity) measured on the lateral wall of the heart of the subject; and Comprising a step of evaluating the left ventricular diastolic function of the subject based on at least one of the first and second ratios; method.

6. In paragraph 2, The step of acquiring the above learning data is: A step of obtaining label data included in the learning data by assigning the above ratio as a label value corresponding to the electrocardiogram data; method.

7. In paragraph 6, A step of training a second neural network model to calculate the ratio of the early diastolic mitral annulus blood flow velocity (E velocity) to the early diastolic mitral annulus tissue velocity (e' velocity) of the subject based on the above learning data; method.

8. A computing device for predicting left ventricular diastolic dysfunction based on electrocardiogram data, a processor comprising at least one core; and a memory including program codes executable by the processor; The above processor, Obtaining evaluation results regarding the left ventricular diastolic function of a subject, and labeling the electrocardiogram data of the subject based on the evaluation results regarding the left ventricular diastolic function of the subject to obtain learning data. Computing device.

9. A computer program stored in a computer-readable storage medium, wherein the computer program, when executed on one or more processors, performs an operation of predicting left ventricular diastolic dysfunction based on electrocardiogram data. The above action is, An operation for obtaining evaluation results regarding the left ventricular diastolic function of a subject; and An operation of obtaining learning data by labeling electrocardiogram data of the subject based on an evaluation result regarding the left ventricular diastolic function of the subject; including; Computer program.

Images