Electronic device for diagnosing disease of user based on biological signal and control method thereof

Abstract

The invention provides an electronic device and a control method thereof. An electronic device according to an embodiment of the present invention comprises: a communication interface; the memory is used for storing the learned first neural network model and the learned second neural network model; one or more processors that acquire biological data of a user and input the acquired biological data into a learned first neural network model to generate first information about a first disease of the user, and a diagnosis unit that inputs the acquired biological data into a learned second neural network model to generate second information on the first disease of the user, and generates a diagnosis result of the user on the first disease of the first type on the basis of the first information and the second information.

Description

Technical Field

[0001] The present invention provides an electronic device for diagnosing a user's disease based on a biological signal and a control method thereof. More specifically, an electronic device for diagnosing a user's disease using a learned neural network model and a control method thereof are disclosed. Background Art

[0002] Myocardial infarction and ischemic heart disease are the leading causes of sudden death, and the number of patients with myocardial infarction and ischemic heart disease is increasing every year. If myocardial infarction and ischemic heart disease are detected and treated early, death and disabilities can be prevented. Therefore, active research has been conducted on the early diagnosis and prediction of myocardial infarction and ischemic heart disease.

[0003] To diagnose myocardial infarction and ischemic heart disease early, an electrocardiogram is measured, and the measured electrocardiogram signal is presented in the form of a curve graph. Based on the curve graph, it is determined whether the patient's heart has myocardial infarction and ischemic heart disease. This method has been widely used. In particular, due to the development of deep learning technology in recent years, deep learning technology has been applied to the medical field. Therefore, a method of obtaining an electrocardiogram signal analysis result by simply inputting the curve graph of the electrocardiogram signal into a neural network model has attracted more and more attention. Summary of the Invention

[0004] Technical Problem

[0005] The present invention aims to solve the problems existing in the foregoing background art. The object of the present invention is to provide an electronic device for diagnosing a user's disease based on a learned neural network model and the user's biological signal and a control method thereof.

[0006] However, the technical problems to be solved by the present invention are not limited to the foregoing problems, and other problems not mentioned above can be clearly understood from the following description.

[0007] Technical Solution

[0008] To solve the technical problems as described above, an embodiment of the present invention discloses a method executed by an electronic device. The method includes the following steps: obtaining biological data of a user; inputting the obtained biological data into a learned first neural network model to generate first information about a first disease of the user; inputting the obtained biological data into a learned second neural network model to generate second information about the first disease of the user; and generating a diagnosis result of the user for the first type of the first disease based on the first information and the second information.

[0009] Alternatively, the step of generating the diagnosis result may include the following steps: based on the second probability value included in the second information, adjusting the first probability value included in the first information, and generating the diagnosis result of the user regarding the first disease based on the first probability value.

[0010] Alternatively, the step of generating the diagnosis result may include the following steps: if the second probability value is above the reference value, applying the weight value corresponding to the second probability value to the first probability value to adjust the first probability value; and if the second probability value is less than the reference value, maintaining the first probability value.

[0011] Alternatively, the step of generating the diagnosis result may include the following steps: if the first probability value is above the first value, generating a first diagnosis result corresponding to the first disease; if the first probability value is less than the first value and above the second value, generating a second diagnosis result corresponding to the first disease; and if the first probability value is less than the second value, generating a third diagnosis result corresponding to the first disease.

[0012] Alternatively, the first information is information on whether the user has the first disease, and the second information may be information on the first type of the first disease.

[0013] Alternatively, the method may include the following steps: based on the first disease and the type of the diagnosis result, extracting a first neural network model and a second neural network model from a plurality of neural network models.

[0014] Alternatively, the step of generating information on whether the user has the first disease and the step of generating information on the first type may be executed in parallel.

[0015] Alternatively, the first neural network model and the second neural network model are learned based on learning data in which a plurality of labels set according to different criteria are assigned to the same electrocardiogram signal, and the plurality of labels may include a first type label corresponding to whether the user has the first disease and a second type label corresponding to the first type of the first disease.

[0016] Alternatively, the biological data includes an electrocardiogram signal, the first disease includes one of myocardial infarction and ischemic heart disease, and the first type may include one of ST-segment elevation myocardial infarction (STEMI) and non-ST-segment elevation myocardial infarction (NSTEMI).

[0017] One embodiment of the present invention for solving the problems described above discloses an electronic device. The electronic device includes: a communication interface; a memory storing a learned first neural network model and a second neural network model; and one or more processors configured to obtain biological data of a user, input the obtained biological data into the learned first neural network model to generate first information about a first disease of the user, input the obtained biological data into the learned second neural network model to generate second information about the first disease of the user, and generate a diagnosis result of the user for a first type of the first disease based on the first information and the second information.

[0018] According to an embodiment of the present invention, when executed by a processor of an electronic device, a non - volatile computer - readable storage medium stores computer instructions that cause the electronic device to perform operations, the operations including the following steps: obtaining biological data of a user; inputting the obtained biological data into a learned first neural network model to generate first information about a first disease of the user; inputting the obtained biological data into a learned second neural network model to generate second information about the first disease of the user; and generating a diagnosis result of the user for a first type of the first disease based on the first information and the second information.

[0019] Effects of the Invention

[0020] The electronic device according to an embodiment of the present invention enables multiple neural network models to learn so as to generate information on whether a disease is present and information on the type of the disease using the same learning data. By doing so, it is possible to ensure the learning data and reduce the costs and time required for pre - processing.

[0021] Moreover, by simply measuring the biological signals of the user, it is possible to provide information on the user's disease and its type, thereby enabling pre - prediction and early diagnosis of the user's disease. Brief Description of the Drawings

[0022] Figure 1 is an exemplary diagram of an electronic device according to an embodiment of the present invention.

[0023] Figure 2 is a block diagram of an electronic device according to an embodiment of the present invention.

[0024] Figure 3 is a sequence diagram schematically showing a control method of an electronic device according to an embodiment of the present invention.

[0025] Figure 4 is an exemplary diagram showing a learned first neural network model and a second neural network model stored in an electronic device according to an embodiment of the present invention.

[0026] Figure 5a FIG. 0 is an exemplary diagram showing a method of causing a first neural network model 20-1 and a second neural network model 20-2 to learn according to an embodiment of the present invention. Figure 5b FIG. 1 is an exemplary diagram showing a method of causing a plurality of neural network models to learn in order to generate information on whether a user has a specific disease and information on a specific type of the specific disease in the prior art.

[0027] Figure 6 FIG. 2 is an exemplary diagram showing a method of diagnosing a disease of a user by using a first neural network model and a second neural network model according to an embodiment of the present invention.

[0028] Figure 7 FIG. 3 is a detailed configuration diagram of an electronic device according to another embodiment of the present invention. DETAILED DESCRIPTION

[0029] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art (hereinafter referred to as "those skilled in the art") can easily implement the present invention. The embodiments disclosed in the present invention are used to enable those skilled in the art to utilize or implement the content of the present invention. Therefore, various modifications to the embodiments of the present invention will be apparent to those skilled in the art. That is, the present invention can be implemented in various different forms and is not limited to the following embodiments.

[0030] Throughout the specification of the present invention, the same or similar reference numerals denote the same or similar elements. Also, some reference numerals related to parts not relevant to the description of the present invention may be omitted in the drawings for clarity of the present invention.

[0031] The term "or" used in the present invention does not mean an exclusive "or" but means an inclusive "or". That is, if the present invention is not specifically specified or its meaning is unclear in the context of a sentence, "X uses A or B" should be understood to mean one of the natural inclusive substitutions. For example, when the present invention is not specifically specified or its meaning is unclear in the context of a sentence, "X uses A or B" can be interpreted as one of the cases where X uses A, the case where X uses B, or the case where X uses both A and B.

[0032] The term "and / or" used in the present invention should be understood to refer to or include all possible combinations of one or more of the recited related concepts.

[0033] The terms "comprise" and / or "have" used in the present invention should be understood to mean the presence of a specific feature and / or element. However, the terms "comprise" and / or "have" should be understood not to exclude the presence or addition of one or more other features, other elements, and / or combinations thereof.

[0034] If the present invention does not specifically specify or cannot clearly indicate the singular form in the context of the sentence, the singular form should generally be interpreted as including "one or more".

[0035] The term "the Nth (N is a natural number)" used in the present invention can be understood as an expression used to distinguish the elements of the present invention from each other according to a preset criterion such as a functional perspective, a structural perspective, or for convenience of description. For example, elements performing different functions in the present invention can be distinguished as the first element or the second element. However, elements that are substantially the same in the technical spirit of the present invention but need to be distinguished for convenience of description can also be distinguished as the first element or the second element.

[0036] The term "acquire" used in the present invention can be understood not only as meaning receiving data through a wired or wireless communication network with an external device or system, but also as meaning generating data in an on-device form.

[0037] On the other hand, the terms "module" or "unit" used in the present invention can be understood as terms referring to an independent functional unit of processing and 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, a "module" or "unit" can be a unit composed of a single element, or a unit represented by a combination or aggregation of multiple elements. For example, in a narrow sense, a "module" or "unit" can refer to a hardware element of an electronic device or an aggregation thereof, an application program that executes a specific function of software, a processing procedure implemented by executing software, or a set of instructions for executing a program. Moreover, in a broad sense, a "module" or "unit" can refer to the electronic device itself constituting a system or an application program executed in the electronic device. However, the foregoing concepts are merely examples, and the concepts of "module" or "unit" can be defined in various ways within the scope understandable by those of ordinary skill in the art based on the content of the present invention.

[0038] The term "model" used in the present invention can be understood as a system implemented using mathematical concepts and language to solve a specific problem, a set of software units for solving a specific problem, or an abstract model of a processing procedure for solving a specific problem. For example, a neural network "model" can refer to the entire system implemented by a neural network that has the ability to solve problems through learning. At this time, the neural network can optimize the parameters of the connected nodes or neurons through learning and thus have the ability to solve problems. A neural network "model" can include a single neural network or a neural network ensemble composed of multiple neural networks combined together.

[0039] The above description of the terms is to help understand the present invention. Therefore, unless the above terms are clearly recorded as matters restricting the content of the present invention, they are not used in the sense of restricting the technical spirit of the content of the present invention.

[0040] Figure 1 It is an exemplary diagram of an electronic device according to an embodiment of the present invention.

[0041] Please refer to Figure 1 , the electronic device 100 according to an embodiment of the present invention obtains the biological signal 10 of the user from an external electronic device 200 linked to the electronic device 100. Here, the external electronic devices 200-1, 200-2 (hereinafter simply denoted as 200) linked to the electronic device 100 are connected to the electronic device 100 through network communication and can be implemented by various devices capable of performing the function of measuring the biological signal 10 (10-1 and 10-2, hereinafter simply denoted as 10) of the users 1-1, 1-2 (hereinafter simply denoted as 1). For example, the external electronic device 200 can be implemented as an electrocardiogram measuring device, a smart watch, a display device, etc. On the other hand, the external electronic device 200 can also pre-register the external electronic device 200 information (or information of an organization where the external electronic device 200 is configured, etc.) into the electronic device 100 in order to be linked with the electronic device 100.

[0042] The electronic device 100 can generate a diagnostic result of the user based on the user signal obtained from the external electronic device 200. Specifically, the electronic device 100 can analyze the user signal obtained from the external electronic device 200 to grasp the user's health status or identify whether the user has a disease and the type of the disease, etc. In particular, in order to generate the diagnostic result of the user, the electronic device 100 can utilize the learned neural network model 20. As an example, the electronic device 100 can input the obtained biological signal 10 into the learned neural network model 20 and obtain the user's health status information or the user's disease information (for example, information including whether the user has a disease or the type of the disease, etc.) based on the output value of the learned neural network model 20. To this end, the learned neural network model 20 can be pre-learned based on various users' biological signals 10 so as to output the user's health status information or the user's disease information, etc.

[0043] On the other hand, the electronic device 100 can transmit the generated diagnostic result to the external electronic device 200 or send it to the user's terminal device. By this means, the user can grasp the health status information only by measuring the biological signal 10 without the assistance of professional facilities or experts.

[0044] The following Figures 2 to 7 will describe in detail the electronic device 100 according to an embodiment of the present invention.

[0045] Figure 2 is a block diagram of the electronic device 100 according to an embodiment of the present invention.

[0046] The electronic device 100 according to an embodiment of the present invention can be a hardware device that performs comprehensive processing and operation of data or a part of the hardware device, or can also be an operation environment based on software connected by a communication network. For example, the electronic device 100 can be a server (for example, a platform server, etc.) that serves as a main body for performing intensive data processing functions and sharing resources, or can also be a client that shares resources through interaction with the server. Moreover, the electronic device 100 can also be a cloud system that comprehensively processes data through the interaction of multiple servers and multiple clients.

[0047] The foregoing is merely an example related to the type of the electronic device 100, and the type of the electronic device 100 can be configured in various ways within the scope understandable by those skilled in the art based on the content of the present invention. For example, the electronic device 100 can also be implemented as an electronic device 100 that measures the user's biological signal (for example, an electrocardiogram signal, etc.). However, in the following, it is assumed that the electronic device 100 of the present invention is a server and then described.

[0048] Please refer to Figure 2, an electronic device 100 according to an embodiment of the present invention may include a memory 110, a communication interface 120, and one or more processors 130. However, Figure 2 merely by way of example, the electronic device 100 may include other elements for implementing an operating environment. Moreover, the electronic device 100 may also include only a part of the plurality of elements disclosed above.

[0049] The memory 110 according to an embodiment of the present invention can be understood as including a hardware and / or software component unit that stores and manages data processed by the electronic device 100. That is, the memory 110 can store any form of data generated or determined by the processor 130 and any form of data received by the processor 130 through the communication interface 120.

[0050] As an example, the memory 110 may store a plurality of learned neural network models. Each neural network model can be matched with at least one other neural network model according to the type of disease diagnosed and the type of diagnostic result generated, and then stored in the memory 110. At this time, the plurality of matched neural network models can be learned in such a way that different types of information are output based on the same input (e.g., biological signals, etc.). For example, the memory 110 may store a first neural network model and a second neural network model. Here, the first neural network model and the second neural network model can be pre-learned based on the same learning data and then stored in the memory 110. The learned first neural network model and the second neural network model can be learned in such a way that different types of information are generated from the biological signals by analyzing the user's biological signals. At this time, the memory 110 may also store the learning data used for the first neural network model and the second neural network model to learn.

[0051] Moreover, as an example, the memory 110 may include at least one type of storage medium in the form of flash memory type, hard disk type, multimedia card micro type, card memory, random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, or optical disk. Moreover, the memory 110 may also include a database system that controls and manages data according to a preset system. The type of the aforementioned memory 110 is merely an example, and the type of the memory 110 can be defined in various ways within the scope understandable by those of ordinary skill in the art based on the content of the present invention.

[0052] The communication interface 120 according to an embodiment of the present invention can be understood as a component unit that transmits and receives data through any known wired or wireless communication system. The electronic device 100 can transmit and receive various information through the communication interface 120 and an external electronic device 100. As an example, the electronic device 100 can also receive, through the communication interface 120, a biological signal of a user measured by the external electronic device 100, or can also transmit diagnostic result information of the user generated by the electronic device 100 to the external electronic device 100.

[0053] To this end, the communication interface 120 can use 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), fifth generation mobile communication (5G), ultra wide-band wireless communication, ZigBee, radio frequency (RF) communication, wireless local area network (WLAN), wireless fidelity (Wi-Fi), near field communication (NFC), or Bluetooth to send and receive data. The foregoing multiple communication systems are merely an example, and various wired or wireless communication systems for the purpose of data transmission and reception of the communication interface 120 can also be applied other than the foregoing example.

[0054] In one embodiment of the present invention, the processor 130 can be electrically connected to the memory 110 and the communication interface 120 to control the overall operation of the electronic device 100.

[0055] The processor 130 can be understood as a constituent unit including hardware and / or software for performing operations. For example, the processor 130 can read a computer program and perform data processing for machine learning. The processor 130 can process operation processes such as processing of input data for machine learning, extraction of features for machine learning, and error calculation based on backpropagation. The processor 130 for performing data processing as described above can include a central processing unit (CPU), a general purpose graphics processing unit (GPGPU), a tensor processing unit (TPU), an application specific integrated circuit (ASIC), or a field programmable gate array (FPGA), etc. The foregoing types of the processor 130 are merely an example, and the types of the processor 130 can be variously configured within the scope understandable by those of ordinary skill in the art based on the content of the present invention.

[0056] Figure 3It is a sequence diagram schematically showing a method of controlling an electronic device 100 according to an embodiment of the present invention.

[0057] Please refer to Figure 3 , the processor 130 obtains the biological data of user 1 (step S310). Specifically, the processor 130 obtains the biological data of user 1 from an external electronic device 200 registered in the electronic device 100 through the communication interface 120. At this time, information about the external electronic device 200 can be stored in the memory 110 of the electronic device 100.

[0058] As an example, the biological data of user 1 can be an electrocardiogram signal. Specifically, the external electronic device 200 can be an electrocardiogram device including a plurality of electrodes (or, a plurality of patches each containing an electrode) connected to different parts of user 1's body. The external electronic device 200 can generate an electrocardiogram signal (or, electrocardiogram data) corresponding to user 1 based on the electrocardiogram of user 1 measured from each electrode and then send it to the electronic device 100. Moreover, the processor 130 can obtain the electrocardiogram signal generated by the external electronic device 200 through the communication interface 120. To help understand the present invention, the biological data is assumed to be an electrocardiogram signal and described below.

[0059] Please refer to Figure 3 , the processor 130 can input the obtained biological data into the learned first neural network model and generate first information about the disease of user 1 (step S320).

[0060] As an example, the first information can be information about whether the user has a disease. Therefore, the first neural network model can be a model pre-learned in such a way as to generate information about whether user 1 has a disease. Specifically, the first neural network model can be a model pre-learned in the following way, that is, when the electrocardiogram signal of user 1 is input, it judges whether user 1 has a disease based on the electrocardiogram signal and generates information about whether user 1 has a disease.

[0061] In particular, the first neural network model can also be pre-learned in such a way as to only judge whether a specific disease is present. To help understand the present invention, the specific disease is referred to as the first disease below.

[0062] As an example, when the electrocardiogram signal of user 1 is input, the first neural network model can judge whether user 1 has the first disease based on the electrocardiogram signal and generate information about whether the first disease is present with an output value.

[0063] At this time, the information on whether the user has the first disease may include a probability value (or score) of possibly having the first disease. Specifically, when the electrocardiogram signal of user 1 is input into the first neural network model, the processor 130 can obtain the probability value of possibly having the first disease in the form of an output value through the Sigmoid layer or SoftMax layer configured at the output end among the multiple layers included in the first neural network model.

[0064] Moreover, the processor 130 can input the obtained biological data into the learned second neural network model and then generate second information about the diseases of user 1 (step S330).

[0065] As an example, the second information may be information about the type of disease. Therefore, the second neural network model may be a model pre-learned in such a way as to generate information about the type of disease of user 1. Specifically, the second neural network model may be pre-learned as follows: if the electrocardiogram signal of user 1 is input, the type of disease of user 1 is identified based on the electrocardiogram signal and information about the type of disease of user 1 is generated.

[0066] In particular, the second neural network model may be pre-learned in such a way as to generate information about the type of a specific disease. At this time, the specific disease may be the same disease as the disease whose presence or absence is identified by the first neural network model (i.e., the first disease).

[0067] Moreover, the second neural network model may also be pre-learned in such a way as to generate information about a specific type of a specific disease. That is, if there are multiple types of the specific disease, the second neural network model can be pre-learned to identify whether the specific disease of user 1 corresponds to a specific type among the multiple types of the specific disease. In particular, the second neural network model can be pre-learned to identify the specific type of the specific disease corresponding to the characteristics grasped after analyzing the biological signal and mastering the characteristics of the biological signal. To help understand the present invention, the specific type is hereinafter referred to as the first type.

[0068] As an example, when the electrocardiogram signal of user 1 is input, the second neural network model can determine whether the first disease of user 1 corresponds to the first type based on the electrocardiogram signal and generate information about the first type of the first disease in the form of an output value.

[0069] At this time, the information on the first type of the first disease may include the probability value (or, score) that the disease of user 1 may correspond to the first type of the first disease. Specifically, when the electrocardiogram signal of user 1 is input into the second neural network model, the processor 130 can obtain the probability value that the disease of user 1 may correspond to the first type of the first disease in the form of an output value through the Sigmoid layer or SoftMax layer arranged at the output end among the multiple layers included in the second neural network model.

[0070] As an example, the first neural network model and the second neural network model can be implemented as a convolutional neural network (CNN) model, a recurrent neural network (RNN) model, a rectified linear unit (Relu) model, etc.

[0071] On the other hand, the processor 130 can also perform a preprocessing process on the acquired biological signal 10 (for example, an electrocardiogram signal) before inputting the acquired biological signal 10 into the first neural network model and the second neural network model.

[0072] On the other hand, according to an embodiment of the present invention, the memory 110 can store multiple neural network models. Here, the multiple neural network models can be matched with other neural network models according to the type of the disease and the type of the diagnosis result. Therefore, after identifying the type of the disease to be diagnosed and the type of the diagnosis result (or, the type of the disease and the type of the diagnosis result set by the user), the processor 130 can determine the combination of the multiple neural network models corresponding to the identified type of the disease and the type of the diagnosis result. Moreover, the processor 130 can extract the multiple neural network models according to the determined combination and generate a diagnosis result for the specific disease of user 1 by using the extracted multiple neural network models. On the other hand, the combination information of the multiple neural network models corresponding to the type of the disease and the type of the diagnosis result can be stored in the memory 110 in a table form.

[0073] For example, if the processor 130 determines to generate a diagnosis result for the first type of the first disease, it can extract the first neural network model and the second neural network model from the multiple neural network models stored in the memory 110. Moreover, the processor 130 can generate a diagnosis result for the first type of the first disease of user 1 after inputting the acquired biological signal into the extracted first neural network model and the second neural network model respectively.

[0074] Figure 4 It is an exemplary diagram showing the learned first neural network model 20-1 and second neural network model 20-2 stored in the electronic device 100 according to an embodiment of the present invention.

[0075] Please refer to Figure 4 such that the processor 130 can input the acquired electrocardiogram signals into the first neural network model 20-1 and the second neural network model 20-2 respectively. At this time, the processor 130 can obtain a plurality of probability values through each neural network model (the first neural network model 20-1 and the second neural network model 20-2). Specifically, the processor 130 can obtain the probability value that the user 1 has the first disease (hereinafter referred to as the first probability value) through the first neural network model 20-1 and obtain the probability value that the disease of the user 1 belongs to the first type of the first disease (hereinafter referred to as the second probability value) through the second neural network model 20-2.

[0076] In particular, please refer to Figure 4 such that the processor 130 can perform the following steps in parallel: generating information on whether the first disease is present; and generating information on the first type of the first disease. That is, the processor 130 inputs the acquired electrocardiogram signals into the first neural network model 20-1 and the second neural network model 20-2 configured in parallel respectively to obtain the output values of the first neural network model 20-1 and the second neural network model 20-2 (i.e., the first probability value and the second probability value), thereby shortening the time taken to generate information on whether the first disease is present and information on the first type of the first disease.

[0077] Figure 5a FIG. is an exemplary diagram showing a method of training the first neural network model 20-1 and the second neural network model 20-2 according to an embodiment of the present invention. Figure 5b FIG. is an exemplary diagram showing a method of training a plurality of neural network models to generate information on whether a specific disease is present and information on a specific type of the specific disease in the prior art.

[0078] According to an embodiment of the present invention, the first neural network model 20-1 and the second neural network model 20-2 can be trained based on learning data in which a plurality of labels set according to different criteria are assigned to the same electrocardiogram signal.

[0079] Specifically, the first neural network model 20-1 and the second neural network model 20-2 may have been trained with the same learning data. As an example, the learning data may include a plurality of electrocardiogram signals and may also be referred to as a learning data set. The first neural network model 20-1 and the second neural network model 20-2 can be trained respectively based on the same learning data including a plurality of electrocardiogram signals. However, in order to generate different types of information (information on whether the first disease is present and information on the first type of the first disease) respectively, different labels can be assigned to the plurality of electrocardiogram signals included in the learning data.

[0080] Here, different tags are tags set according to different criteria, and may include a first type of tag corresponding to whether or not a first disease is present and a second type of tag corresponding to a first category of the first disease. That is, referring to Figure 5a , the learning data of an embodiment of the present invention can each be given a first type of tag corresponding to whether or not a first disease is present and a second type of tag corresponding to a first category of the first disease.

[0081] Therefore, the first neural network model 20-1 can learn based on the first type of tag of the learning data in order to generate information about whether or not a first disease is present, and the second neural network model 20-2 can learn based on the second type of tag of the learning data in order to generate information about a first category of the first disease.

[0082] Referring to Figure 5b , in the prior art, in order to generate information about whether or not a specific disease is present and information about a specific category of a specific disease, it is necessary to prepare independent learning data for each neural network model to learn. For this reason, a lot of cost and time are required for preprocessing each learning data. In particular, for learning data obtained by a professional device such as the biological signal 10 (for example, an electrocardiogram signal, etc.), it is difficult to prepare sufficient learning data for each neural network model to learn.

[0083] However, according to an embodiment of the present invention, in order to generate different types of information (information about whether or not a first disease is present and information about a first category of the first disease), the same learning data given multiple tags set according to different criteria is used, so that the learning data can be reused to allow multiple neural network models (that is, the first neural network model 20-1 and the second neural network model 20-2) to learn. By this, the cost and time required for preprocessing the learning data can be reduced, and sufficient learning data for each neural network model to learn can be ensured.

[0084] However, the present invention is not limited to the foregoing embodiment. That is, the first neural network model 20-1 and the second neural network model 20-2 can also learn with different learning data. Specifically, the first neural network model 20-1 can learn as follows, that is, based on first learning data including electrocardiogram data given a first tag (a tag for distinguishing myocardial infarction), identify whether or not myocardial infarction is present, and the second neural network model 20-2 can learn as follows, that is, based on second learning data including electrocardiogram data given a second tag (a tag for distinguishing ST-segment elevation myocardial infarction), identify whether or not it is equivalent to ST-segment elevation myocardial infarction.

[0085] Please refer back to Figure 3, the processor 130 can generate a diagnosis result of the user 1 for the first type of the first disease based on the information on whether the user has the first disease and the information on the first category of the first disease (step S340).

[0086] Specifically, the processor 130 can combine the information on whether the user has the first disease and the information on the first category to generate a diagnosis result of the first disease for the user 1. As an example, the processor 130 can combine the information on whether the user has the first disease and the information on the first category of the first disease to generate a diagnosis result of the first type of the first disease. That is, when using the first neural network model 20-1, the processor 130 can only determine whether the user has the first disease. However, according to an embodiment of the present invention, the processor 130 can combine the result of the first neural network model 20-1 (i.e., whether the user has the first disease) and the result of the second neural network model 20-2 (i.e., whether it belongs to the first category of the first disease) to generate a diagnosis result of the first type of the first disease.

[0087] On the other hand, the processor 130 can also send the generated diagnosis result of the first disease to the external electronic device 200 through the communication interface 120 to provide it to the user 1.

[0088] On the other hand, according to an embodiment of the present invention, the first disease includes one of myocardial infarction and ischemic heart disease, and the first category of the first disease can be a high-risk group of myocardial infarction. That is, the processor 130 can generate information on whether the user 1 has myocardial infarction or ischemic heart disease based on the electrocardiogram signal through the first neural network model 20-1, and can generate information on the high-risk group of myocardial infarction for the user 1 through the second neural network model 20-2. For the convenience of explaining the present invention, the first disease is assumed to be myocardial infarction for the following description.

[0089] As an example, when the processor 130 identifies one of ST elevation myocardial infarction (STEMI) or non-ST elevation myocardial infarction (NSTEMI) based on the electrocardiogram signal of the user 1, it determines that the myocardial infarction of the user 1 belongs to the high-risk group.

[0090] Therefore, the second neural network model 20-2 can learn in the following manner, that is, based on the input electrocardiogram signal, calculate whether user 1 is equivalent to ST-segment elevation myocardial infarction or non-ST-segment elevation myocardial infarction and use it as the output value. At this time, the second probability value can be the probability value that user 1's acute myocardial infarction is equivalent to ST-segment elevation myocardial infarction or the probability value that it is equivalent to non-ST-segment elevation myocardial infarction. That is, the processor 130 can judge that user 1's disease is equivalent to one of ST-segment elevation myocardial infarction or non-ST-segment elevation myocardial infarction based on the second probability value obtained through the second neural network model 20-2. On the other hand, considering that both ST-segment elevation myocardial infarction and non-ST-segment elevation myocardial infarction are all types of acute myocardial infarction, the information output through the second neural network model 20-2 (for example, the second probability value) can be referred to as information about the first type of the first disease. That is, the second neural network model 20-2 can be pre-learned to output information about the first type of the first disease.

[0091] For the convenience of explaining the present invention, hereinafter, the first type of the first disease is assumed to be ST-segment elevation myocardial infarction for explanation.

[0092] On the other hand, according to an embodiment of the present invention, the first type of the first disease can be acute myocardial infarction. Therefore, the processor 130 can be as follows, that is, judge that user 1 has myocardial infarction based on the first probability value, and judge that ST-segment elevation myocardial infarction of user 1 is recognized based on the second probability value, and then generate a diagnosis result about user 1's acute myocardial infarction.

[0093] Hereinafter, an embodiment of the present invention for generating a diagnosis result of user 1 based on the first probability value and the second probability value will be described in detail.

[0094] Figure 6 It is a sequence diagram showing a method for diagnosing the disease of user 1 using the first neural network model 20-1 and the second neural network model 20-2 according to an embodiment of the present invention. Figure 6 The steps S610 to S630 shown can each correspond to Figure 3 the steps S310 to S330 shown, and the detailed description thereof will be omitted hereinafter.

[0095] According to an embodiment of the present invention, the processor 130 can adjust the first probability value included in the first information based on the second probability value included in the second information, and generate a diagnosis result of user 1 about the first type of the first disease based on the first probability value.

[0096] Specifically, if the processor 130 identifies the first type of the first disease based on the output value (i.e., the second probability value) of the second neural network model, the first probability value regarding whether the user has the first disease can be adjusted based on the second probability value.

[0097] As an example, if the first type of the first disease is identified through the second neural network model, it means that the probability that the user has the first type of the first disease is relatively high. Therefore, the processor 130 can adjust the first probability value regarding whether the user has the first disease based on the second probability value regarding whether the user has the first type of the first disease.

[0098] As an example, when the second probability value is above the reference value, the processor 130 can apply the weight value corresponding to the second probability value to the first probability value to adjust the first probability value. Moreover, when the second probability value is less than the reference value, the processor 130 can keep the first probability value.

[0099] Specifically, the processor 130 can determine whether the second probability value is above the reference value. Here, the reference value can be set differently for each user 1 according to the age, body type, disease history, etc. of the user 1.

[0100] As an example, the processor 130 can determine that the second probability value is above the reference value and the user 1 has the first type of the first disease. Specifically, the processor 130 can make the following determination, that is, the second probability value is above the reference value and ST-segment elevation myocardial infarction is identified based on the electrocardiogram signal of the user 1. Then, when the processor 130 determines that ST-segment elevation myocardial infarction can be identified, it can apply the weight value corresponding to the second probability value to the first probability value to adjust the first probability value. On the contrary, the processor 130 can make the following determination. That is, the second probability value is less than the reference value and ST-segment elevation myocardial infarction cannot be identified based on the electrocardiogram signal of the user 1. Then, when the processor 130 determines that ST-segment elevation myocardial infarction cannot be identified, it can keep the first probability value without applying the weight value corresponding to the second probability value.

[0101] More specifically, when the first probability value is 40, the second probability value is 32, and the reference value is 30, the processor can recognize that the second probability value is above the reference value and determine that the user has ST-segment elevation myocardial infarction. At this time, the processor 130 can determine the weight value corresponding to the second probability value. For example, if the weight value is set to 0.5 times the second probability value (32), then the processor 130 can apply the weight value (16 = 32 × 0.5) to the first probability value (40) to obtain the adjusted first probability value (56 = 32 + 16). On the contrary, when the first probability value is 40, the second probability value is 24, and the reference value is 30, the processor can recognize that the second probability value is less than the reference value and determine that the user does not have ST-segment elevation myocardial infarction. At this time, the processor 130 can keep the obtained first probability value (40) unchanged.

[0102] On the other hand, the processor 130 can adjust the first probability value based on the second probability value by various methods. As an example, if the second probability value is above the reference value, the first probability value can be changed to the second probability value.

[0103] Or, the processor 130 can also apply different weight values according to the difference between the first probability value and the second probability value. Specifically, it can be applied as follows: if the second probability value is above the reference value and the difference between the first probability value and the second probability value is above a preset difference value, the first weight value is applied; if the second probability value is above the reference value and the difference between the first probability value and the second probability value is less than the preset difference value, the second weight value smaller than the first weight value is applied.

[0104] Or, the processor 130 can also apply different weight values according to the section in which the first probability value is included among a plurality of preset sections. Specifically, it can be applied as follows: if the second probability value is above the reference value and the first probability value is included in the first section (0 or more and less than the first reference value), the third weight value is applied; if the second probability value is above the reference value and the first probability value is included in the second section (the first reference value or more and less than the second reference value), the fourth weight value smaller than the third weight value is applied; if the first probability value is included in the third section (the second reference value or more and less than the third reference value), the fifth weight value smaller than the fourth weight value is applied.

[0105] On the other hand, the processor 130 can generate a diagnosis result based on the first probability value. Here, the first probability value may include the first probability value adjusted based on the second probability value or the first probability value output by the first neural network model (i.e., the first probability value that is kept).

[0106] As an example, the processor 130 may be as follows. That is, if the first probability value is greater than or equal to the first value, a first diagnosis result corresponding to the first disease of the first type is generated; if the first probability value is less than the first value and greater than or equal to the second value, a second diagnosis result corresponding to the first disease of the first type is generated; and if the first probability value is less than the second value, a third diagnosis result corresponding to the first disease of the first type is generated.

[0107] More specifically, if the first value is 3 and the second value is set to 48.5, the processor 130 may generate a diagnosis result that the user is normal (or a diagnosis result with a low probability of acute myocardial infarction) when the first probability value (the adjusted first probability value or the maintained first probability value) is less than the first value (3). Moreover, the processor 130 may generate a diagnosis result that the user belongs to the medium-risk group of acute myocardial infarction (or a diagnosis result with a certain probability of acute myocardial infarction) when the first probability value is greater than or equal to the first value (3) and less than the second value (48.5). Finally, the processor 130 may generate a diagnosis result that the user belongs to the high-risk group of myocardial infarction (or a diagnosis result with a high probability of acute myocardial infarction) when the first probability value is greater than or equal to the second value (48.5).

[0108] On the other hand, the processor 130 may make a judgment as follows. That is, the first probability value is compared with the second probability value, the first value, and the second value, and if it is recognized that the first probability value is less than the first value, it is judged that the user 1 does not have a myocardial infarction disease; if it is recognized that the second probability value is greater than or equal to the reference value, it is judged that the myocardial infarction of the user 1 belongs to the high-risk group. At this time, the processor 130 may recognize the electrocardiogram signal of the user 1 as abnormal, generate information requesting the re-measurement of the electrocardiogram signal of the user 1, and send it to the external electronic device 200 through the communication interface 120 or output it through an output interface (such as a speaker, a display, etc.).

[0109] According to an embodiment of the present invention, the electronic device 100 can be implemented as an electrocardiogram measuring device. The following is combined with Figure 7 to illustrate the electronic device implemented as an electrocardiogram measuring device according to an embodiment of the present invention.

[0110] Figure 7 It is a specific configuration diagram of an electronic device according to another embodiment of the present invention.

[0111] At this time, the electronic device 100' includes a memory 110', a communication interface 120', a measurement unit 140', a display 150', a user interface 160', and a processor 130'. For Figure 7 the shown elements, the memory 110', the communication interface 120', and the processor 130', the same can be applied Figure 2Descriptions of the memory 110, communication interface 120, and processor 130 shown are provided, so their detailed descriptions will be omitted below.

[0112] The measurement unit 140’ includes a plurality of leads connected to different parts of the user 1's body, and can generate an electrocardiogram signal of the user 1 based on the voltage differences between the body parts of the user 1 measured through the plurality of leads. For example, the measurement unit 140’ may include limb leads and precordial leads. At this time, the limb leads may include 4 electrodes connected to the four limbs (hereinafter referred to as "limb electrodes"). Moreover, the precordial leads may include 6 electrodes connected to the chest (hereinafter referred to as "chest electrodes").

[0113] The limb electrodes may include a right arm electrode RA, a left arm electrode LA, a right leg electrode RL, and a left leg electrode LL. The right leg electrode RL may be a common electrode or a ground electrode. The limb electrodes may be respectively connected to the positions corresponding to the right arm, left arm, right leg, and left leg.

[0114] Moreover, the chest electrodes (or, precordial chest electrodes) may include a first chest electrode V1, a second chest electrode V2, a third chest electrode V3, a fourth chest electrode V4, a fifth chest electrode V5, and a sixth chest electrode V6.

[0115] The display 150’ can output various types of image information. For example, the processor 130 may output the diagnosis result of the user 1 through the display 150’. Or, the processor 130 may also output a first type of warning information for the first disease of the user 1 through the display 150’.

[0116] For this purpose, the display 150’ can be implemented as a display including a self-emitting device or a display including a non-self-emitting device and a backlight. For example, it can be implemented as various types of displays such as a liquid crystal display panel (LCD), an organic light emitting diode (OLED) display, a light emitting diode (LED), a micro liquid crystal display panel (micro LED), a mini liquid crystal display panel (Mini LED), a plasma display panel (PDP), a quantum dot (QD) display, and a quantum dot light-emitting diode (QLED).

[0117] The display 150' may also include a driving circuit, a backlight unit, etc. implemented in forms such as amorphous silicon thin film transistors (a-si TFTs), low temperature polycrystalline silicon thin film transistors (LTPS, low temperature poly silicon TFTs), and organic thin film transistors (OTFTs, organic TFTs). On the other hand, the display 150' can be implemented as a touch screen combined with a touch sensor, a flexible display, a rollable display, a 3D display, a display physically connected with multiple display modules, etc.

[0118] Moreover, the display 150' can also form a touch screen together with a touch panel.

[0119] The user interface 160' is an element used when interacting with the electronic device 100 and the user 1, and the processor 130 can output a diagnosis result through the user interface 160'. On the other hand, the user interface 160' may include at least one of a touch sensor, a motion sensor, a button, a jog dial, a switch, and a microphone, but the present invention is not limited thereto.

[0120] The various embodiments of the present invention described above can be combined with additional embodiments and can be modified within the scope understandable by those skilled in the art according to the foregoing specific description. It should be understood that the embodiments of the present invention are illustrative rather than restrictive in all aspects. For example, each structural element described in a single form can be implemented dispersedly, and similarly, the structural elements described in a dispersed form can also be implemented in a combined form. Therefore, all changes or modified forms derived from the meaning, scope, and equivalent concepts of the claims of the present invention should be construed as falling within the scope of the present invention.

Claims

1. A method, which is executed by an electronic device including at least one processor, comprising the following steps: Obtain the biological data of the user; Input the obtained biological data into a learned first neural network model to generate first information about a first disease of the user; Input the obtained biological data into a learned second neural network model to generate second information about the first disease of the user; And Based on the first information and the second information, generate a diagnosis result of the user for the first type of the first disease.

2. The method according to claim 1, characterized in that, The step of generating the diagnosis result includes the following steps: Based on the second probability value included in the second information, adjust the first probability value included in the first information, and based on the first probability value, generate a diagnosis result of the user for the first type of the first disease.

3. The method according to claim 2, wherein The step of generating the diagnosis result includes the following steps: If the second probability value is above a reference value, apply a weight value corresponding to the second probability value to the first probability value to adjust the first probability value; and If the second probability value is less than the reference value, keep the first probability value.

4. The method according to claim 2, wherein The step of generating the diagnosis result includes the following steps: If the first probability value is above a first value, generate a first diagnosis result corresponding to the first type of the first disease; if the first probability value is less than the first value and above a second value, generate a second diagnosis result corresponding to the first type of the first disease; if the first probability value is less than the second value, generate a third diagnosis result corresponding to the first type of the first disease.

5. The method according to claim 1, wherein The first information is information about whether the user has the first disease, and the second information is information about the first category of the first disease.

6. The method according to claim 1, characterized in that, Including the following steps: Based on the first disease and the type of the diagnosis result, extract the first neural network model and the second neural network model from multiple neural network models.

7. The method according to claim 5, characterized in that The biological data includes an electrocardiogram signal, The first disease includes one of myocardial infarction and ischemic heart disease, The first category includes one of ST-segment elevation myocardial infarction and non-ST-segment elevation myocardial infarction.

8. The method according to claim 1, characterized in that The first neural network model and the second neural network model are learned based on learning data in which multiple labels set according to different criteria are assigned to the same electrocardiogram signal, The multiple labels include a first type label corresponding to whether the user has the first disease and a second type label corresponding to the first category of the first disease.

9. The method according to claim 1, characterized in that, The step of generating the first information about the first disease and the step of generating the second information about the first disease are executed in parallel.

10. An electronic device, characterized in that Comprising: A communication interface; A memory storing a learned first neural network model and a second neural network model; One or more processors, obtain the biological data of the user through the communication interface, input the obtained biological data into the learned first neural network model to generate first information about the first disease of the user, input the obtained biological data into the learned second neural network model to generate second information about the first disease of the user, and generate a diagnosis result of the user for the first disease of the first type based on the first information and the second information.

11. A method, executed by an electronic device including at least one processor, comprising the following steps: Obtain the biological data of the user; Input the obtained biological data into the learned first neural network model to generate first information about the first disease of the user; Input the obtained biological data into the learned second neural network model to generate second information about the first disease of the user; And Generate a diagnosis result of the user for the first disease of the first type based on the first information and the second information.

Images