Method for generating disease prediction information using biometric signal

Abstract

According to an embodiment of the present disclosure, disclosed is a method for generating disease prediction information using biometric signals, which is executed by a computing device. The method may comprise the steps of: acquiring biometric signals at a diagnosis time point; acquiring past biometric signals at a previous time point before the diagnosis time point or information associated with the past biometric signals; and generating disease prediction information for the diagnosis time point on the basis of the biometric signals at the diagnosis time point and the past biometric signals at the past time point or the information associated with the past biometric signals, using a neural network model.

Description

A method for generating disease prediction information using biosignals

[0001] The present invention relates to a method for generating disease prediction information, and more particularly, to a technology for generating disease prediction information at the time of diagnosis by utilizing bio-signals and additional information at the time of diagnosis.

[0002] With the advancement of medical measurement technology, individuals can easily and inexpensively measure their biosignals by visiting medical institutions such as hospitals, and with the advancement of information and communication technology, the measured biosignals can be easily managed through database management, etc.

[0003] Additionally, recently, various bio-signals can be acquired relatively easily and inexpensively through portable measuring devices or wearable devices without individuals having to visit medical institutions, and the history of acquired bio-signals can be easily managed through various data management systems.

[0004] Meanwhile, existing AI models for diagnosing specific diseases mostly predicted the probability of disease onset based on a single biosignal measured at the time of diagnosis. In other words, existing AI models mostly predicted diseases based on biosignals at a single point in time (e.g., biosignals at the current time of diagnosis). Therefore, existing AI models could not comprehensively consider biosignals from various points in time or consider biosignal history, and because they predicted diseases based only on fragmented biosignals at a single point in time, there was a problem of relatively low prediction accuracy.

[0005] Korean Patent No. 10-0748184 (August 3, 2007) discloses a device and method for diagnosing heart disease based on electrocardiogram using a neural network.

[0006] The present disclosure aims to provide a technology for more accurately predicting disease prediction information at the time of diagnosis by utilizing past bio-signals from past points in time in addition to bio-signals at the time of diagnosis.

[0007] Meanwhile, the technical task to be achieved by the present disclosure is not limited to the technical task mentioned above, and may include various technical tasks within a scope obvious to a person skilled in the art from the contents described below.

[0008] According to one embodiment of the present disclosure for realizing the task described above, a method for generating disease prediction information using a biosignal performed by a computing device is disclosed. The method may include the steps of: obtaining a biosignal at a time of diagnosis; obtaining a past biosignal at a time in the past before the time of diagnosis or information associated with the past biosignal; and using a neural network model, generating disease prediction information for the time of diagnosis based on the biosignal at the time of diagnosis and the past biosignal at the time of the past or information associated with the past biosignal.

[0009] In one embodiment, the step of generating disease prediction information for the time of diagnosis may include: generating a first vector based on the biosignal; generating a second vector based on the past biosignal or information associated with the past biosignal; and generating the disease prediction information for the time of diagnosis based on the first vector and the second vector by utilizing the neural network model.

[0010] In one embodiment, the information associated with the past biosignal may include at least one of information regarding a time difference between the past point in time and the diagnosis point in time; medical information of the past point in time; or reference prediction information generated for the past biosignal using a reference neural network model different from the neural network model.

[0011] In one embodiment, the medical information of the past point in time may include medical measurement information measured within a predetermined time from the time of acquisition of the past biosignal.

[0012] In one embodiment, the reference neural network model may be a model that generates a reference prediction value based on a biosignal at a single point in time.

[0013] In one embodiment, the step of generating the disease prediction information for the diagnosis time point based on the first vector and the second vector by utilizing the neural network model may include the step of generating an integrated vector based on the first vector and the second vector; and the step of generating the disease prediction information for the diagnosis time point based on the integrated vector by utilizing the neural network model.

[0014] In one embodiment, the first vector and the second vector are vectors generated through different preprocessing methods, the first vector and the second vector include different numbers of elements, and the integrated vector may include more elements of the first vector than elements of the second vector.

[0015] In one embodiment, the step of obtaining a past biosignal or information associated with the past biosignal for a past time point prior to the diagnosis time point may include a step of obtaining a plurality of past biosignals or information associated with the plurality of past biosignals for a plurality of past time points prior to the diagnosis time point, and the step of generating a second vector based on the past biosignal or information associated with the past biosignal may include a step of generating the second vector based on information associated with the plurality of past biosignals.

[0016] In one embodiment, the information associated with the plurality of past bio-signals includes respective individual information associated with each past bio-signal included in the plurality of past bio-signals, and the respective individual information associated with each past bio-signal may include at least one of information among: time difference information between each past point in time and the diagnosis point in time; medical information mapped to each past bio-signal; or reference prediction information generated for each past bio-signal using a reference neural network model different from the neural network model.

[0017] In one embodiment, the step of generating the second vector based on information associated with the plurality of past bio-signals may include, when the number of the plurality of past bio-signals is less than a predetermined number, the step of generating the insufficient number of supplementary bio-signals by utilizing at least one past bio-signal among the plurality of past bio-signals; and the step of generating the second vector based on the plurality of past bio-signals and information associated with the supplementary bio-signals.

[0018] In one embodiment, the supplementary biosignals may be generated based on a past biosignal from the most recent point in time among the plurality of past biosignals.

[0019] In one embodiment, the information associated with the plurality of past bio-signals includes tabular data generated based on the plurality of past bio-signals, and the tabular data may include at least one type of data among a first type of data related to time differences; a second type of data related to medical information; a third type of data related to disease prediction based on a single point in time; or a fourth type of data related to missing information.

[0020] According to one embodiment of the present disclosure for realizing the task described above, a structure of a neural network model implemented by a computing device is disclosed. The structure of the neural network model may include a preprocessing module for preprocessing a biosignal at a time of diagnosis and preprocessing a past biosignal at a time in the past before the time of diagnosis or information associated with the past biosignal; and a prediction module for generating disease prediction information based on data related to the biosignal generated by the preprocessing module and data related to the past biosignal generated by the preprocessing module.

[0021] In one embodiment, the preprocessing module may include a first encoder module that generates a first vector based on the biosignal at the time of diagnosis; a second encoder module that generates a second vector based on the past biosignal or information associated with the past biosignal; and an integration module that generates an integrated vector based on the first vector and the second vector.

[0022] In one embodiment, the prediction module can generate the disease prediction information based on the integrated vector.

[0023] According to one embodiment of the present disclosure for achieving the above-described task, a computer program stored in a computer-readable storage medium is disclosed. When the computer program is executed on one or more processors, the computer program causes the one or more processors to perform the following operations for generating disease prediction information using a biosignal, wherein the operations may include: an operation of acquiring a biosignal at a time of diagnosis; an operation of acquiring a past biosignal at a time in the past before the time of diagnosis or information associated with the past biosignal; and an operation of generating disease prediction information for a time of diagnosis based on the biosignal at the time of diagnosis and the past biosignal at the time of diagnosis or information associated with the past biosignal by utilizing a neural network model.

[0024] A computing device according to one embodiment of the present disclosure for achieving the above-described task is disclosed. The device includes at least one processor; and a memory, wherein the at least one processor is configured to acquire a biosignal at a time of diagnosis; acquire a past biosignal at a past time prior to the time of diagnosis or information associated with the past biosignal; and generate disease prediction information for the time of diagnosis based on the biosignal at the time of diagnosis and the past biosignal at the past time or information associated with the past biosignal by utilizing a neural network model.

[0025] The present disclosure can further improve the accuracy of a neural network model's disease prediction by utilizing past biosignals in addition to biosignals at the time of diagnosis. For example, the present disclosure can implement a neural network model that learns a complex correlation between biosignals at the time of diagnosis, past biosignals, and diseases. Utilizing this neural network model can further improve the accuracy of disease prediction.

[0026] Furthermore, the present disclosure can further improve the accuracy of disease prediction by utilizing a neural network model that considers multifaceted information related to past biosignals. For example, the present disclosure can utilize multifaceted secondary processed information related to past biosignals as input to a neural network model, allowing the neural model to learn multifaceted correlations related to past biosignals. Through these operations, the accuracy of disease prediction can be further improved.

[0027] Furthermore, the present disclosure can further improve the accuracy of disease prediction by utilizing a neural network model that considers information on multiple past biosignals from multiple past time points in addition to the biosignals at the time of diagnosis. For example, the present disclosure can implement a neural network model that more deeply reflects the correlation between biosignals' histories, and such a neural network model can further improve the accuracy of disease prediction.

[0028] Meanwhile, the effects of the present disclosure are not limited to the effects mentioned above, and various effects may be included within a range apparent to those skilled in the art from the contents described below.

[0029] FIG. 1 is a block diagram of a computing device for generating disease prediction information using biosignals according to one embodiment of the present disclosure.

[0030] FIG. 2 is a conceptual diagram illustrating a neural network according to one embodiment of the present disclosure.

[0031] FIG. 3 is a flowchart illustrating a method for generating disease prediction information using biosignals according to one embodiment of the present disclosure.

[0032] FIG. 4 is a diagram illustrating an embodiment of information associated with a single past biosignal according to one embodiment of the present disclosure.

[0033] FIG. 5 is a diagram for explaining an operation of generating reference prediction information for past biosignals by utilizing a reference neural network model different from a neural network model according to one embodiment of the present disclosure.

[0034] FIG. 6 is a diagram for explaining information associated with a plurality of past biosignals according to one embodiment of the present disclosure.

[0035] FIG. 7 is a diagram for explaining an operation of generating a second vector when the number of multiple past bio-signals satisfies a preset number according to one embodiment of the present disclosure.

[0036] FIG. 8 is a diagram for explaining an operation of generating a second vector when the number of multiple past biosignals according to one embodiment of the present disclosure is less than a predetermined number.

[0037] FIG. 9 is a block diagram of the structure of a neural network model for generating disease prediction information using biosignals according to one embodiment of the present disclosure.

[0038] FIG. 10 is a diagram schematically illustrating an operation of generating disease prediction information in a neural network model according to one embodiment of the present disclosure.

[0039] FIG. 11 is a simplified, general schematic diagram of an exemplary computing environment in which embodiments of the present disclosure may be implemented.

[0040] Various embodiments are now described with reference to the drawings. In this specification, various descriptions are provided to facilitate understanding of the present disclosure. However, it will be apparent that these embodiments may be practiced without these specific details.

[0041] As used herein, the terms "component," "module," "system," and the like refer to computer-related entities, hardware, firmware, software, a combination of software and hardware, or an execution of software. For example, a component may be, but is not limited to, a procedure running on a processor, a processor, an object, a thread of execution, a program, and / or a computer. For example, both an application running on a computing device and the computing device may be a component. One or more components may reside within a processor and / or a thread of execution. A component may be localized within a single computer. A component may be distributed between two or more computers. Furthermore, these components may execute from various computer-readable media having various data structures stored therein. Components may communicate via local and / or remote processes, for example, by signals comprising one or more data packets (e.g., data from one component interacting with another component in a local system, a distributed system, and / or data transmitted to another system via a network such as the Internet via signals).

[0042] Furthermore, the term "or" is intended to mean an inclusive "or" rather than an exclusive "or." That is, unless otherwise specified or clear from context, "X employs A or B" is intended to mean either of the natural inclusive permutations. That is, if X employs A; X employs B; or X employs both A and B, "X employs A or B" can apply to any of these cases. Furthermore, the term "and / or" as used herein should be understood to refer to and include all possible combinations of one or more of the associated items listed.

[0043] Additionally, the terms "comprises" and / or "comprising" should be understood to imply the presence of the features and / or components in question. 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, components, and / or groups thereof. Furthermore, unless otherwise specified or clear from the context to refer to the singular form, the singular in the specification and claims should generally be construed to mean "one or more."

[0044] And, the term "at least one of A or B" should be interpreted to mean "if it includes only A", "if it includes only B", or "if it is combined in the composition of A and B".

[0045] Those skilled in the art should further appreciate that the various illustrative logical blocks, configurations, modules, circuits, means, logics, and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both. To clearly illustrate the interchangeability of hardware and software, various illustrative components, blocks, configurations, means, logics, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application. However, such implementation decisions should not be interpreted as causing a departure from the scope of the present disclosure.

[0046] The description of the disclosed embodiments is provided to enable a person skilled in the art to make or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art. The general principles defined herein may be applied to other embodiments without departing from the scope of the present disclosure. Therefore, the present invention is not limited to the embodiments disclosed herein. The present invention is to be construed in the widest scope consistent with the principles and novel features disclosed herein.

[0047] In the present disclosure, network function and artificial neural network and neural network can be used interchangeably.

[0048]

[0049] FIG. 1 is a block diagram of a computing device for generating disease prediction information using biosignals according to one embodiment of the present disclosure.

[0050] The configuration of the computing device (100) illustrated in FIG. 1 is merely a simplified example. In one embodiment of the present disclosure, the computing device (100) may include other configurations for performing the computing environment of the computing device (100), and only some of the disclosed configurations may constitute the computing device (100).

[0051] A computing device (100) may include a processor (110), memory (130), and network unit (150).

[0052] The processor (110) may be configured with one or more cores, and may include a processor for data analysis and deep learning, such as a central processing unit (CPU), a general purpose graphics processing unit (GPGPU), and a tensor processing unit (TPU) of a computing device. The processor (110) may read a computer program stored in the memory (130) to perform data processing for machine learning according to an embodiment of the present disclosure. According to an embodiment of the present disclosure, the processor (110) may perform operations for learning a neural network. The processor (110) may perform calculations for learning a neural network, such as processing input data for learning in deep learning (DL), extracting features from input data, calculating errors, and updating weights of a neural network using backpropagation. At least one of the CPU, GPGPU, and TPU of the processor (110) may process learning of a network function. For example, a CPU and a GPGPU can jointly process network function learning and data classification using network functions. Furthermore, in one embodiment of the present disclosure, processors of multiple computing devices can be jointly used to process network function learning and data classification using network functions. Furthermore, a computer program executed on a computing device according to one embodiment of the present disclosure may be a CPU, GPGPU, or TPU executable program.

[0053] According to one embodiment of the present disclosure, the memory (130) can store any form of information generated or determined by the processor (110) and any form of information received by the network unit (150).

[0054] According to one embodiment of the present disclosure, the memory (130) 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 (e.g., SD or XD memory, etc.), 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. The computing device (100) may also operate in relation to web storage that performs the storage function of the memory (130) on the internet. The description of the above-described memory is merely an example, and the present disclosure is not limited thereto.

[0055] The network unit (150) according to one embodiment of the present disclosure can use various wired communication systems such as a public switched telephone network (PSTN), xDSL (x Digital Subscriber Line), RADSL (Rate Adaptive DSL), MDSL (Multi Rate DSL), VDSL (Very High Speed ​​DSL), UADSL (Universal Asymmetric DSL), HDSL (High Bit Rate DSL), and a local area network (LAN).

[0056] In addition, the network unit (150) presented in this specification can use various wireless communication systems such as CDMA (Code Division Multi Access), TDMA (Time Division Multi Access), FDMA (Frequency Division Multi Access), OFDMA (Orthogonal Frequency Division Multi Access), SC-FDMA (Single Carrier-FDMA) and other systems.

[0057] In the present disclosure, the network unit (150) may be configured regardless of the communication mode, such as wired or wireless, and may be configured as various communication networks, such as a local area network (LAN), a personal area network (PAN), and a wide area network (WAN). In addition, the network may be the well-known World Wide Web (WWW), and may also utilize a wireless transmission technology used for short-distance communication, such as infrared (IrDA: Infrared Data Association) or Bluetooth.

[0058] The techniques described in this specification can be used in other networks as well as the networks mentioned above.

[0059]

[0060] FIG. 2 is a conceptual diagram illustrating a neural network according to one embodiment of the present disclosure.

[0061] Throughout this specification, the terms computational model, neural network, network function, and neural network may be used interchangeably. A neural network may be comprised of a set of interconnected computational units, generally referred to as nodes. These nodes may also be referred to as neurons. A neural network comprises at least one node. The nodes (or neurons) comprising a neural network may be interconnected by one or more links.

[0062] Within a neural network, one or more nodes connected via links can form a relationship between input nodes and output nodes. The concept of input nodes and output nodes is relative, meaning that any node that is in an output node relationship with one node can also be in an input node relationship with another node, and vice versa. As described above, the relationship between input nodes and output nodes can be created based on links. One input node can be connected to one or more output nodes via links, and vice versa.

[0063] In a relationship between input nodes and output nodes connected through a single link, the data of the output node can have its value determined based on the data input to the input node. Here, the link interconnecting the input nodes and output nodes can have a weight. The weight can be variable and can be varied by the user or an algorithm so that the neural network can perform a desired function. For example, when one or more input nodes are interconnected to one output node through each link, the output node can determine the output node value based on the values ​​input to the input nodes connected to the output node and the weight set on the link corresponding to each input node.

[0064] As described above, a neural network is a network in which one or more nodes are interconnected through one or more links, forming input and output node relationships within the network. The characteristics of a neural network can be determined based on the number of nodes and links within the network, the relationships between the nodes and links, and the weights assigned to each link. For example, if two neural networks have the same number of nodes and links but different weight values ​​for the links, the two neural networks can be perceived as different from each other.

[0065] A neural network can be composed of a set of one or more nodes. A subset of the nodes comprising the neural network can form a layer. Some of the nodes comprising the neural network can form a layer based on their distances from the initial input node. For example, a set of nodes that are n distances from the initial input node can form n layers. The distance from the initial input node can be defined by the minimum number of links required to reach the node from the initial input node. However, this definition of a layer is arbitrary for illustrative purposes, and the degree of a layer within a neural network can be defined in a different way than described above. For example, a layer of nodes can be defined by its distance from the final output node.

[0066] An initial input node may refer to one or more nodes within a neural network into which data is directly input without going through links with other nodes. Alternatively, within a neural network, it may refer to nodes that do not have other input nodes connected by links in the relationship between nodes based on links. Similarly, a final output node may refer to one or more nodes within a neural network that do not have output nodes in their relationship with other nodes. Furthermore, a hidden node may refer to nodes that constitute a neural network other than the initial input node and the final output node.

[0067] A neural network according to one embodiment of the present disclosure may be a neural network in which the number of nodes in an input layer may be the same as the number of nodes in an output layer, and the number of nodes decreases and then increases as it progresses from the input layer to the hidden layer. In addition, a neural network according to another embodiment of the present disclosure may be a neural network in which the number of nodes in an input layer may be less than the number of nodes in an output layer, and the number of nodes decreases as it progresses from the input layer to the hidden layer. In addition, a neural network according to another embodiment of the present disclosure may be a neural network in which the number of nodes in an input layer may be greater than the number of nodes in an output layer, and the number of nodes increases as it progresses from the input layer to the hidden layer. A neural network according to another embodiment of the present disclosure may be a neural network in a combined form of the neural networks described above.

[0068] A deep neural network (DNN) can refer to a neural network that includes multiple hidden layers in addition to input and output layers. Using DNNs, one can identify latent structures in data. For example, one can identify the latent structures of images, text, videos, audio, and music (e.g., what objects are in the image, what the content and emotion of the text are, what the content and emotion of the audio are, etc.). DNNs can include convolutional neural networks (CNNs), recurrent neural networks (RNNs), autoencoders, generative adversarial networks (GANs), restricted Boltzmann machines (RBMs), deep belief networks (DBNs), Q networks, U networks, Siamese networks, and generative adversarial networks (GANs). The description of the deep neural network described above is only an example and the present disclosure is not limited thereto.

[0069] In one embodiment of the present disclosure, the network function may include an autoencoder. An autoencoder may be a type of artificial neural network that outputs output data similar to input data. The autoencoder may include at least one hidden layer, and an odd number of hidden layers may be arranged between input and output layers. The number of nodes in each layer may be reduced from the number of nodes in the input layer to an intermediate layer called a bottleneck layer (encoding), and then expanded symmetrically from the bottleneck layer to the output layer (symmetrical to the input layer). The autoencoder may perform nonlinear dimensionality reduction. The number of input layers and output layers may correspond to the dimensionality after preprocessing of the input data. In the autoencoder structure, the number of nodes in the hidden layer included in the encoder may have a structure in which the number of nodes decreases as it moves away from the input layer. The number of nodes in the bottleneck layer (the layer with the fewest nodes between the encoder and decoder) may be kept above a certain number (e.g., more than half of the input layer), as too few nodes may not transmit enough information.

[0070] Neural networks can learn through at least one of the following methods: supervised learning, self-supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning. Learning a neural network can be the process of applying knowledge to the neural network to perform a specific action.

[0071] Neural networks can be trained to minimize output errors. This process involves repeatedly inputting training data into the neural network, calculating the neural network output and target error for the training data, and backpropagating the neural network error from the output layer to the input layer to update the weights of each node in the neural network to reduce the error. Supervised learning uses training data with the correct answer labeled for each training data (i.e., labeled training data). Unsupervised learning, on the other hand, may not have the correct answer labeled for each training data. For example, in the case of supervised learning for data classification, the training data may be data with each category labeled. Labeled training data is input to the neural network, and the error can be calculated by comparing the output (category) of the neural network with the training data labels. Alternatively, in the case of unsupervised learning for data classification, the error can be calculated by comparing the input training data with the neural network output. The calculated error is backpropagated in the neural network in the backward direction (i.e., from the output layer to the input layer), and the connection weights of each node in each layer of the neural network can be updated according to the backpropagation. The amount of change in the connection weights of each node to be updated can be determined by the learning rate. The neural network's calculation of the input data and the backpropagation of the error can constitute a learning cycle (epoch). The learning rate can be applied differently depending on the number of iterations of the neural network's learning cycle. For example, a high learning rate can be used in the early stages of neural network training to quickly achieve a certain level of performance, thereby increasing efficiency. A lower learning rate can be used in the later stages of training to increase accuracy.

[0072] In neural network training, training data can typically be a subset of real-world data (i.e., the data to be processed using the trained neural network). Therefore, there can be a learning cycle where errors on the training data decrease but errors on the real-world data increase. Overfitting is a phenomenon where excessive training on the training data leads to increased errors on the real-world data. For example, a neural network trained on yellow cats may fail to recognize cats when shown non-yellow colors, a type of overfitting. Overfitting can increase errors in machine learning algorithms. Various optimization methods can be used to prevent overfitting. These methods include increasing the training data, regularization, dropout, which disables some nodes in the network during the learning process, and the use of batch normalization layers.

[0073]

[0074] The present disclosure proposes a model learning method utilizing past biosignals and information associated with past biosignals to improve the performance of a model that generates disease prediction information utilizing a neural network model.

[0075] The present disclosure effectively enhances the performance of a model that generates disease prediction information at the time of diagnosis by maximizing the use of past biosignals measured by an individual patient at a point in time prior to diagnosis and medical data associated with those past biosignals. Furthermore, the present disclosure exhibits improved performance compared to models that utilize only a single biosignal at the time of diagnosis, even when medical data associated with past biosignals is unavailable.

[0076]

[0077] FIG. 3 is a flowchart illustrating a method for generating disease prediction information using a biosignal according to one embodiment of the present disclosure, FIG. 4 is a diagram for explaining an embodiment of information associated with a single past biosignal according to one embodiment of the present disclosure, and FIG. 5 is a diagram for explaining an operation of generating reference prediction information for a past biosignal using a reference neural network model different from a neural network model according to one embodiment of the present disclosure.

[0078]

[0079] The bio-signals mentioned below may include various bio-related signals such as electrocardiogram (ECG), electromyogram (EMG), electroencephalogram (EEG), electrodermal activity (EDA), skin temperature (SKT), photoplethysmography (PPG), vital signs, X-ray image signals, magnetic resonance imaging (MRI) signals, and computed tomography (CT) image signals. In addition, the bio-signals are not limited to these examples, and may include various information obtained from a subject using a medical device, a wearable device, a mobile device, etc.

[0080] Meanwhile, the following describes embodiments of the present disclosure, focusing on an embodiment in which the bio-signal is an electrocardiogram (ECG) signal. For example, the present disclosure is described focusing on an embodiment in which a neural network model is used to generate disease prediction information at the time of diagnosis based on the bio-signal at the time of diagnosis (the ECG signal at the time of diagnosis) and past bio-signals (the past ECG signals). However, the above-described matters are merely examples, and the present disclosure is not limited thereto.

[0081]

[0082] First, we would like to explain the operation of obtaining “past bio-signals of a single past point in time” from a computing device (100) through FIGS. 3 to 5 and generating disease prediction information for a diagnosis point in time.

[0083] For reference, the "past point in time" described below means a point in time that occurred before the time of diagnosis, and the units for the past point in time can be categorized into, for example, seconds, minutes, hours, days, weeks, months, and years.

[0084] Additionally, the diagnosis time point and past time points may include macroscopic or microscopic temporal information related to the measurement of biosignals. For example, if the unit of time point is seconds, minutes, or hours, the diagnosis time point and past time points may include microscopic temporal information indicating the start point, midpoint, and end point of the measurement of biosignals. Additionally, if the unit of time point is days, weeks, months, or years, the diagnosis time point and past time points may include macroscopic temporal information including the measurement duration of the biosignals.

[0085]

[0086] Referring to FIG. 3 according to one embodiment of the present disclosure, a method for generating disease prediction information using a biosignal may include a step of obtaining a biosignal at the time of diagnosis (S110), a step of obtaining a past biosignal at a past time before the time of diagnosis (S120), and a step of generating disease prediction information for the time of diagnosis based on the biosignal at the time of diagnosis and the past biosignal at the past time by using a neural network model (S130). The method for generating disease prediction information using a biosignal illustrated in FIG. 3 may be performed by a computing device (100).

[0087]

[0088] According to one embodiment of the present disclosure, the computing device (100) can obtain a bio-signal at the time of diagnosis (current time) (S110). For example, the computing device (100) can obtain the bio-signal at the time of diagnosis through medical equipment installed in a medical institution, a portable measuring device (e.g., a smart watch, a patch, a bar-type electrocardiogram measuring device, etc.). Alternatively, the computing device (100) can also obtain the bio-signal at the time of diagnosis from an electronic medical record (EMR), etc.

[0089] In addition, the computing device (100) can obtain information (10) associated with the bio-signal at the time of diagnosis. For example, referring to FIG. 4, the computing device (100) can obtain, in addition to the bio-signal at the time of diagnosis, unique ID (PID) information associated with the bio-signal at the time of diagnosis, date information for the time of diagnosis, etc. In addition, the "PID" illustrated in FIG. 4 may be information associated (related) with the patient's unique identification information. However, the information (10) associated with the bio-signal at the time of diagnosis is not limited thereto, and may include information associated with various bio-signals.

[0090]

[0091] According to one embodiment of the present disclosure, the computing device (100) can acquire past bio-signals from a past point in time prior to the diagnosis time (S120). In one embodiment, the computing device (100) can acquire past bio-signals from a past point in time prior to the diagnosis time through a medical institution's database, a storage of a portable measuring device, etc. Alternatively, the computing device (100) can also acquire past bio-signals from a past point in time prior to the diagnosis time from an electronic medical record (EMR), etc.

[0092] In addition, the computing device (100) can obtain information (20) associated with the past bio-signal of the past time point. For example, referring to FIG. 4, the computing device (100) can obtain, in addition to the past bio-signal of the past time point, PID information associated with the past bio-signal of the past time point, date information for the past time point, etc. In addition, the “PID” illustrated in FIG. 4 may be information associated with (related to) patient-specific identification information. However, the information (20) associated with the past bio-signal of the past time point is not limited thereto, and may include information associated with various bio-signals.

[0093] In one embodiment, the information (20) associated with the past biosignal may further include at least one of information of a time difference (21) between the past point in time and the diagnosis point in time, medical information (22) mapped to the past biosignal, or reference prediction information (23) generated for the past biosignal using a reference neural network model different from the neural network model.

[0094] For example, the time difference information (21) may be information calculated by calculating the time difference between the past point in time and the diagnosis point in time. For example, referring to FIG. 4, the computing device (100) may obtain the time difference information (21) based on the difference between the first time included in the information (20) associated with the past biosignal and the second time included in the information (10) associated with the biosignal at the diagnosis point in time. For example, if the first time is July 13, 2008 and the second time is April 27, 2016, the computing device (100) may obtain the time difference information of 7.7927 from the time difference between the second time and the first time. For reference, the unit of the time difference information here may be a year.

[0095] In addition, the computing device (100) may obtain medical information (22) mapped to the past bio-signal from an electronic medical record (EMR). However, the present invention is not limited thereto, and the computing device (100) may obtain medical information mapped to the past bio-signal from various methods. For example, the medical information (22) may include medical measurement information measured within a predetermined time from the time of acquisition of the past bio-signal. For example, if the neural network model that generates disease prediction information for the time of diagnosis is a model that diagnoses heart failure, the past bio-signal may be an electrocardiogram (ECG) signal, and the medical measurement information may include a left ventricular ejection fraction (EF) value measured within 15 days of the measurement interval between the past acquired ECG signal and the heart failure diagnosis. In addition, if the neural network model that generates disease prediction information at the time of diagnosis is a model that diagnoses estimated glomerular filtration rate (eGFR) and abnormal blood potassium concentration, the past biosignal may be an electrocardiogram (ECG) signal, and the medical measurement information may include the ECG signal and blood test values ​​with a measurement interval of less than 1 hour. However, the present invention is not limited thereto, and various embodiments may exist. Meanwhile, the computing device (100) may improve the performance of the neural network model that diagnoses the biosignal at the time of diagnosis by additionally obtaining medical information that is mapped to the past biosignal. For example, the computing device (100) may implement a neural network model that improves the accuracy of disease prediction information by comprehensively considering additional medical information related to the past biosignal during the prediction process.

[0096] According to one embodiment, the computing device (100) may generate reference prediction information (23) for the past biosignal by utilizing a reference neural network model different from the neural network model. For reference, the neural network model is a model that generates disease prediction information for the time of diagnosis based on the biosignal at the time of diagnosis and the past biosignals at past time points (i.e., a model that generates disease prediction information for the time of diagnosis by utilizing biosignals of multiple time points including the time of diagnosis). In addition, the reference neural network model may be a separate model that generates a reference prediction value based on a biosignal at a single time point. For example, referring to FIG. 5, the computing device (100) may obtain a biosignal at a single time point (e.g., an electrocardiogram signal). In addition, the computing device (100) may perform preprocessing on the biosignal at the single time point (e.g., an electrocardiogram signal). Thereafter, the computing device (100) may input the preprocessed biosignal at the single time point into the reference neural network model to diagnose a specific disease. According to one embodiment, the reference neural network model may be a model trained to diagnose a case where the left ventricular ejection fraction (EF) is 40% or less from a single-time point bio-signal acquired using 12-lead. In addition, the reference neural network model may output a prediction result of a specific disease as a value greater than or equal to 0 and less than or equal to 1. According to one embodiment of the present disclosure, a "separate neural network model based on a single time point (i.e., a separate neural network model that performs a diagnosis by analyzing a single-time point bio-signal)" is used, and separate prediction information based on a single time point generated for the past bio-signal itself is additionally included in the information (20) associated with the past bio-signal and may be considered by the neural network model, so that more complex information related to the past bio-signal may affect the prediction of the neural network model. Therefore, through this operation, the accuracy of the prediction of the neural network model may be further improved.

[0097]

[0098] According to one embodiment of the present disclosure, the computing device (100) can generate disease prediction information for the time of diagnosis based on the bio-signal at the time of diagnosis and the past bio-signal at the past time by utilizing a neural network model (S130).

[0099] Here, the neural network model can perform an operation (three optional operations) based on the biosignal at the time of diagnosis, or based on information (10) associated with the biosignal at the time of diagnosis, or based on the biosignal at the time of diagnosis and information (10) associated with the biosignal at the time of diagnosis, and in parallel perform an operation (three optional operations) based on the past biosignal, or based on information (20) associated with the past biosignal, or based on the past biosignal and information (20) associated with the past biosignal, and finally generate disease prediction information for the time of diagnosis based on combinations of nine optional operations. Meanwhile, the biosignal at the time of diagnosis, the information associated with the biosignal at the time of diagnosis (10), the past biosignal, or the information associated with the past biosignal (20) may be input to the neural network model and then embedded and vectorized, or may be input to the neural network model in a vectorized state by preprocessing.

[0100] In one embodiment, the computing device (100) may i) generate a first vector based on the bio-signal at the time of diagnosis. For example, the computing device (100) may generate the first vector by utilizing a feature extraction model. In other words, the computing device (100) may input the bio-signal at the time of diagnosis into the feature extraction model to generate the first vector (e.g., a feature vector). For example, the feature extraction model may be implemented as various types of models, such as a fully connected layer, a CNN, an RNN, and a transformer.

[0101] In addition, the computing device (100) may ii) generate a second vector based on information associated with the past biosignal. For example, referring to FIG. 4, the computing device (100) may generate a second vector based on information (20) associated with the past biosignal. In one embodiment, the computing device (100) may generate tabular data based on the information (20) associated with the past biosignal, and may generate the second vector based on the tabular data. For example, the tabular data may include, as shown in FIG. 4, a first type of data (21) associated with a time difference, a second type of data (22) associated with medical information, a third type of data (23) associated with disease prediction based on a single point in time, etc.

[0102] In addition, the computing device (100) may iii) generate the disease prediction information for the diagnosis time point based on the first vector and the second vector by utilizing the neural network model. For example, the computing device (100) may generate an integrated vector based on the first vector and the second vector. In addition, the computing device (100) may generate the disease prediction information for the diagnosis time point based on the integrated vector by utilizing the neural network model. Here, the first vector and the second vector include different numbers of elements, and the integrated vector may include more elements of the first vector than elements of the second vector. Since the prediction information of the neural network model is ultimately for the diagnosis time point, and information about the past time point is merely auxiliary information for increasing the accuracy of prediction for the diagnosis time point, it is preferable that the integrated vector include more elements of the first vector than elements of the second vector.

[0103] Meanwhile, the operation of generating the disease prediction information for the diagnosis time point based on the first vector and the second vector by utilizing the neural network model will be described in more detail through FIG. 9 and FIG. 10 below.

[0104]

[0105] Next, with reference to FIGS. 3, 6, and 8, an operation of acquiring “multiple past bio-signals” from a computing device (100) and generating disease prediction information for a diagnosis time point will be described.

[0106] FIG. 6 is a diagram for explaining information associated with a plurality of past biosignals according to one embodiment of the present disclosure.

[0107] Referring again to FIG. 3 according to one embodiment of the present disclosure, a method for generating disease prediction information using a biosignal may include a step of acquiring a biosignal at the time of diagnosis (S110), a step of acquiring a past biosignal at a past time before the time of diagnosis (S120), a step of generating disease prediction information for the time of diagnosis based on the biosignal at the time of diagnosis and the past biosignal at the past time using a neural network model (S130), etc. The method for generating disease prediction information using a biosignal illustrated in FIG. 3 may be performed by a computing device (100).

[0108] Meanwhile, in the embodiments discussed below, in relation to step S120, multiple past biosignals from multiple past time points (rather than a single past biosignal from a single time point) may be acquired. Furthermore, in step S130, the multiple past biosignals may be considered in the process of generating the disease prediction information.

[0109]

[0110] According to one embodiment of the present disclosure, the computing device (100) can acquire a biosignal at the time of diagnosis (S110). For example, the computing device (100) can acquire the biosignal at the time of diagnosis through medical equipment installed in a medical institution, a portable measuring device (e.g., a smartwatch, a patch, a bar-type electrocardiogram measuring device, etc.). Alternatively, the computing device (100) can also acquire the biosignal at the time of diagnosis from an electronic medical record (EMR), etc.

[0111] In addition, the computing device (100) can obtain information (10) associated with the bio-signal at the time of diagnosis. For example, referring to FIG. 6, the computing device (100) can obtain, in addition to the bio-signal at the time of diagnosis, unique ID (PID) information associated with the bio-signal at the time of diagnosis, date information for the time of diagnosis, etc. Meanwhile, in a medical record, “PID” may be information associated with (related to) the patient’s unique identification information. However, the information (10) associated with the bio-signal at the time of diagnosis is not limited thereto, and may include information associated with various bio-signals.

[0112]

[0113] According to one embodiment of the present disclosure, the computing device (100) can obtain past bio-signals of past points in time prior to the diagnosis time (S120). At this time, the computing device (100) can obtain "a plurality of past bio-signals for a plurality of past points in time" prior to the diagnosis time. In one embodiment, the computing device (100) can obtain a plurality of past bio-signals for a plurality of past points in time prior to the diagnosis time through a database of a medical institution, a storage of a portable measuring device, etc. Alternatively, the computing device (100) can also obtain a plurality of past bio-signals for a plurality of past points in time prior to the diagnosis time from an electronic medical record (EMR), etc.

[0114] In addition, the computing device (100) can obtain information (201) associated with a plurality of past bio-signals of the plurality of past points in time. For example, referring to FIG. 6, the computing device (100) can obtain, in addition to the plurality of past bio-signals of the plurality of past points in time, PID information associated with each past bio-signal, date information for each past point in time, etc. In addition, the “PID” illustrated in FIG. 6 may be information associated with (related to) patient-specific identification information. However, the information (201) associated with the plurality of past bio-signals is not limited thereto and may include information associated with various bio-signals.

[0115] In one embodiment, the information (201) associated with the plurality of past bio-signals may include each additional individual information associated with each past bio-signal included in the plurality of past bio-signals. The each additional individual information associated with each past bio-signal may include at least one of information of a time difference (21) between each past point in time and the diagnosis point in time, medical information (22) mapped to each past bio-signal, or reference prediction information (23) generated for each past bio-signal using a reference neural network model different from the neural network model.

[0116] Here, the time difference information (21) may be time difference information between each of the past points in time and the diagnosis point in time. For example, referring to FIG. 6, the computing device (100) may obtain the time difference information (21) based on the difference between each time included in the information (201) associated with the plurality of past bio-signals and the time included in the information (10) associated with the bio-signal at the diagnosis point in time. For example, if the first time included in the information associated with the bio-signal at the diagnosis point in time is April 27, 2016, and the second time included in the information (201) associated with the plurality of past bio-signals is July 13, 2008, the computing device (100) may obtain time difference information of 7.7927 from the time difference between the second time and the first time. Through the same method, if the third time included in the information (201) associated with the plurality of past bio-signals is January 14, 2012, time difference information of 4.3148 can be obtained from the time difference between the third time and the first time. The computing device (100) can generate a plurality of time difference information between the plurality of past time points included in the information (201) associated with the plurality of past bio-signals and the diagnosis time point.

[0117] In addition, the computing device (100) may obtain each piece of medical information (22) mapped to each of the past bio-signals from an electronic medical record (EMR). However, the present invention is not limited thereto, and the computing device (100) may obtain each piece of medical information mapped to each of the past bio-signals through various methods. For example, the medical information (22) may include each piece of medical measurement information measured within a predetermined time from the acquisition time of each of the past bio-signals. For example, if the neural network model that generates disease prediction information for the time of diagnosis is a model that diagnoses heart failure, each piece of past bio-signal may be an electrocardiogram (ECG) signal, and each piece of medical measurement information may include a left ventricular ejection fraction (EF) value measured within 15 days of the measurement interval between the past acquired ECG signal and the heart failure diagnosis or information on whether or not heart failure has been diagnosed. In addition, if the neural network model that generates disease prediction information at the time of diagnosis is a model that diagnoses estimated glomerular filtration rate (eGFR) and abnormal blood potassium concentration, each past biosignal may be an electrocardiogram (ECG) signal, and each medical measurement information may include an ECG signal and a blood test value with a measurement interval of less than 1 hour. However, the present invention is not limited thereto, and various embodiments may exist. Meanwhile, the computing device (100) can improve the performance of the neural network model that diagnoses the biosignal at the time of diagnosis by additionally obtaining each medical information mapped to each past biosignal. For example, the computing device (100) can implement a neural network model that improves the accuracy of disease prediction information by comprehensively considering additional medical information related to the past biosignal during the prediction process.

[0118] According to one embodiment, the computing device (100) may generate reference prediction information (23) for each of the past biosignals by utilizing a reference neural network model different from the neural network model. For reference, the neural network model is a model that generates disease prediction information for the time of diagnosis based on the biosignal at the time of diagnosis and a plurality of past biosignals of a plurality of past time points. In addition, the reference neural network model may be a separate model that generates a reference prediction value based on a biosignal at a single time point. For example, referring again to FIG. 5, the computing device (100) may obtain a biosignal at a single time point (e.g., an electrocardiogram signal). In addition, the computing device (100) may perform preprocessing on the biosignal at the single time point (e.g., an electrocardiogram signal). Thereafter, the computing device (100) may input the preprocessed biosignal at the single time point into the reference neural network model to diagnose a specific disease. According to one embodiment, the reference neural network model may be a model trained to diagnose a case where the left ventricular ejection fraction (EF) is 40% or less from a single-time point bio-signal acquired using 12-lead. The reference neural network model may output a prediction result of a specific disease as a value greater than or equal to 0 and less than or equal to 1. As described above, according to one embodiment of the present disclosure, a plurality of separate prediction pieces of information based on a single time point, generated for each past bio-signal itself using a "separate neural network model based on a single time point," may be additionally included in the information (201) associated with the plurality of past bio-signals and may be considered by the neural network model, so that more complex information related to the plurality of past bio-signals may affect the prediction of the neural network model. Therefore, through this operation, the accuracy of the prediction of the neural network model may be further improved.

[0119]

[0120] According to one embodiment of the present disclosure, the computing device (100) can generate disease prediction information for the time of diagnosis based on the biosignal at the time of diagnosis and the past biosignals of the past time points by utilizing the neural network model (S130). In particular, the computing device (100) can generate disease prediction information for the time of diagnosis based on the biosignal at the time of diagnosis and the plurality of past biosignals of the plurality of past time points by utilizing the neural network model.

[0121] Here, the neural network model can perform an operation (three optional operations) based on the bio-signal at the time of diagnosis, or based on information (10) associated with the bio-signal at the time of diagnosis, or based on the bio-signal at the time of diagnosis and information (10) associated with the bio-signal at the time of diagnosis, and in parallel perform an operation (three optional operations) based on the plurality of past bio-signals, or based on information (201) associated with the plurality of past bio-signals, or based on the plurality of past bio-signals and information (20) associated with the plurality of past bio-signals, and finally generate disease prediction information for the time of diagnosis based on combinations of nine optional operations. Meanwhile, the bio-signal at the time of diagnosis, information (10) associated with the bio-signal at the time of diagnosis, the plurality of past bio-signals, or information (201) associated with the plurality of past bio-signals may be input to the neural network model and then embedded and vectorized, or may be input to the neural network model in a vectorized state through preprocessing.

[0122] In one embodiment, the computing device (100) may i) generate a first vector based on the bio-signal at the time of diagnosis. For example, the computing device (100) may generate the first vector by utilizing a feature extraction model. In other words, the computing device (100) may input the bio-signal at the time of diagnosis into the feature extraction model to generate the first vector (e.g., a feature vector). For example, the feature extraction model may be implemented as various types of models such as a fully connected layer, a CNN, an RNN, and a transformer.

[0123] In addition, the computing device (100) may ii) generate a second vector based on information associated with the plurality of past bio-signals. For example, referring to FIG. 6, the computing device (100) may generate the second vector based on information (201) associated with the plurality of past bio-signals. In one embodiment, the computing device (100) may generate tabular data based on the information (201) associated with the plurality of past bio-signals, and may generate the second vector based on the tabular data. For example, the tabular data may include, as shown in FIG. 6, a first type of data (21) associated with a time difference, a second type of data (22) associated with medical information, a third type of data (23) associated with disease prediction based on a single time point, etc.

[0124]

[0125] According to one embodiment of the present disclosure, the neural network model may include a network that performs disease prediction by utilizing information associated with a biosignal at the time of diagnosis and a predetermined number (e.g., 5) of past biosignals.

[0126] In such cases, for the neural network model to perform normal prediction operations, a predetermined number of past bio-signals must be secured. However, in some situations (e.g., situations where past measurement history is insufficient, situations where measurements are taken by a new patient, situations where database loss occurs, etc.), it may be difficult to secure the predetermined number of past bio-signals. Therefore, even in such situations where past bio-signals are missing, a missing information processing operation is required to ensure normal operation of the neural network model.

[0127] In one embodiment, when it is difficult to secure the predetermined number of past biosignals, the computing device (100) may generate supplementary biosignals to supplement the insufficient number of missing signals. For example, the computing device (100) may generate the insufficient number of supplementary biosignals by utilizing at least one acquired past biosignal. Specifically, the computing device (100) may generate the insufficient number of supplementary biosignals by directly copying at least one acquired past biosignal and information associated therewith, generate the insufficient number of supplementary biosignals by combining two or more acquired past biosignals, or generate the insufficient number of supplementary biosignals through various methods other than these methods.

[0128] Meanwhile, the computing device (100) may generate the insufficient number of supplementary biosignals based on the most recent past biosignal among the acquired past biosignals. For example, the computing device (100) may generate the insufficient number of supplementary biosignals by copying the most recent past biosignal and its associated information. When performing the supplementary operation in this manner, the computation is simplified, so that computational resources can be managed efficiently, and the neural network model can be prevented from performing predictions by excessively concentrating on information with low temporal distribution near the time of diagnosis.

[0129] Additionally, the computing device (100) may add data related to missing information to the information (201) associated with the plurality of biosignals. For example, after generating the insufficient number of supplementary biosignals, the computing device (100) may add data indicating that each supplementary biosignal was supplementarily generated according to a missing situation to the information (201) associated with the plurality of biosignals. In one embodiment, the information (201) associated with the plurality of biosignals may be implemented as the tabular data described above, and in this case, the tabular data may additionally include a fourth type of data related to missing information. Here, the fourth type of data may indicate whether the past biosignal is an actually acquired past biosignal or corresponds to a supplementary biosignal.

[0130]

[0131] FIG. 7 is a diagram illustrating an operation of generating a second vector when the number of multiple past biosignals satisfies a preset number according to one embodiment of the present disclosure. In other words, the embodiment of FIG. 7 is an example showing a situation in which the neural network model operates depending on a predetermined number of past biosignals and in which all of the predetermined number of past biosignals are acquired (i.e., a situation in which there are no missing values).

[0132] Referring to FIG. 7, the computing device (100) may acquire a predetermined number of past bio-signals (e.g., 5) and generate tabular data (30) based on the acquired past bio-signals. Here, the tabular data (30) may include information associated with the predetermined number of past bio-signals. In addition, the computing device (100) may generate the second vector for the predetermined number of past bio-signals based on the tabular data (30).

[0133] In addition, the computing device (100) can generate an integrated vector based on the first vector and the second vector by utilizing the neural network model, and generate the disease prediction information for the diagnosis time point based on the integrated vector. Here, the first vector and the second vector include different numbers of elements, and the integrated vector can include more elements of the first vector than elements of the second vector. That is, the neural network model can generate an integrated vector by placing more weight on a bio-signal at the diagnosis time point than on a plurality of past bio-signals of a plurality of past time points, and can predict a disease by utilizing this integrated vector.

[0134] For example, referring to FIG. 7, the first vector may be generated based on a bio-signal at a diagnosis time corresponding to April 27, 2016. In addition, the second vector may be generated based on information associated with each of a predetermined number of past bio-signals acquired at a predetermined number (e.g., 5) of past time points (e.g., September 10, 2014, January 21, 2013, January 20, 2012, January 4, 2012, July 13, 2008) prior to April 27, 2016 (the diagnosis time). For example, the computing device (100) may generate tabular data (30) based on the predetermined number of past bio-signals, and generate the second vector based on the tabular data (30). Meanwhile, the operation of generating the disease prediction information for the diagnosis time point based on the first vector and the second vector by utilizing the neural network model will be described in more detail with reference to FIGS. 9 and 10 below.

[0135]

[0136] FIG. 8 is a diagram illustrating an operation of generating a second vector when the number of multiple past biosignals is less than a predetermined number according to one embodiment of the present disclosure. In other words, the embodiment of FIG. 8 is an example that illustrates a situation in which the neural network model operates depending on a predetermined number of past biosignals and the number of acquired past biosignals is less than the predetermined number (i.e., a situation in which there are missing values).

[0137] Referring to FIG. 8, the computing device (100) may generate the insufficient number of supplementary biosignals by utilizing at least one past biosignal among the acquired past biosignals when “the number of acquired past biosignals is less than a predetermined number (e.g., 5).” For example, the supplementary biosignals may be generated based on the past biosignal of the most recent time point among the acquired past biosignals. In addition, the computing device (100) may generate the second vector based on information associated with the acquired past biosignals and information associated with the supplementary biosignals. For example, the computing device (100) may generate the second vector based on data (30) in a table format, such as FIG. 8, including information associated with the acquired past biosignals and information associated with the supplementary biosignals.

[0138] In addition, the computing device (100) can generate an integrated vector based on the first vector and the second vector by utilizing the neural network model, and generate the disease prediction information for the diagnosis time point based on the integrated vector. Here, the first vector and the second vector include different numbers of elements, and the integrated vector can include more elements of the first vector than elements of the second vector. That is, the neural network model can generate an integrated vector by placing more weight on a bio-signal at the diagnosis time point than on a plurality of past bio-signals of a plurality of past time points, and can predict a disease by utilizing this integrated vector.

[0139] For example, referring to FIG. 8, the first vector may be generated based on a bio-signal at a diagnosis time corresponding to April 27, 2016. In addition, the second vector may be generated based on information associated with two past bio-signals acquired at two past points in time (January 4, 2012, July 13, 2008) prior to April 27, 2016 (the diagnosis time), when the predetermined number is 5, and information associated with three supplementary bio-signals generated for supplementation. In addition, the information associated with the three supplementary bio-signals may all be generated based on a past bio-signal acquired at July 13, 2008, which is the most past point in time among the two acquired past bio-signals. In one embodiment, the information associated with the three supplementary bio-signals may be generated by directly copying the first type of data (e.g., time difference information), the second type of data (e.g., medical information), and the third type of data of the past bio-signals acquired on July 13, 2008, which is the most recent point in time, and setting the fourth type of data to an opposite value. Meanwhile, the operation of generating the disease prediction information for the diagnosis time point based on the first vector and the second vector by utilizing the neural network model will be described in more detail with reference to FIGS. 9 and 10 below.

[0140]

[0141] Hereinafter, with reference to FIGS. 9 and 10, the operation of generating the disease prediction information for the diagnosis time point based on the first vector and the second vector by utilizing the neural network model will be described in more detail.

[0142] FIG. 9 is a block diagram of a structure of a neural network model for generating disease prediction information using a biosignal according to an embodiment of the present disclosure, and FIG. 10 is a diagram schematically illustrating an operation of generating disease prediction information in a neural network model according to an embodiment of the present disclosure. For reference, in FIG. 10, the biosignal at the time of diagnosis obtained and the past biosignal at a time in the past before the time of diagnosis may be an electrocardiogram (ECG) signal. In addition, the neural network model may be a model that diagnoses a case where the left ventricular ejection fraction (EF) is 40% or less using an ECG signal. However, the present invention is not limited thereto, and the neural network model may be implemented in various ways as a model that generates specific disease prediction information in response to various biosignals that are obtained.

[0143]

[0144] Referring to FIG. 9, the structure of the neural network model may include a preprocessing module (200) and a prediction module (300).

[0145] According to one embodiment of the present disclosure, the preprocessing module (200) can preprocess bio-signals at the time of diagnosis and past bio-signals from past points in time prior to the time of diagnosis. For example, the preprocessing module (200) may include a first encoder module (210), a second encoder module (220), and an integration module (230).

[0146] According to one embodiment, the first encoder module (210) may generate a first vector based on the bio-signal at the time of diagnosis. For example, referring to FIG. 10, the bio-signal at the time of diagnosis may be an electrocardiogram (ECG) signal measured on April 27, 2016. In one embodiment, the first encoder module (210) may generate the first vector by utilizing various neural network architectures such as SE-ResNeXt 152 and CoAtNet after receiving the preprocessed ECG signal at the time of diagnosis. Alternatively, the first encoder module (210) may generate the first vector by utilizing various neural network architectures such as a fully connected layer, CNN, RNN, and transformer. Meanwhile, in the example of FIG. 10, the first vector may be implemented in the form of a vector including 2048 elements, but the number of elements of the first vector is not limited to this example, and may be configured in various ways under the condition of including more elements than the elements of the second vector.

[0147] In addition, the second encoder module (220) can generate a second vector based on the past bio-signals. For example, the second encoder module (220) can generate the second vector based on information associated with a plurality of past bio-signals. For example, referring to FIG. 10, the second encoder module (220) can obtain tabular data (30) generated based on the plurality of past bio-signals. Here, the tabular data (30) can include a first type of data related to a time difference, a second type of data related to medical information, a third type of data related to disease prediction based on a single time point, a fourth type of data related to missing information, etc. For example, the tabular data (30) may be generated based on the acquired past bio-signals, as described above, i) when the number of acquired past bio-signals satisfies a predetermined number (e.g., 5), or ii) when the number of acquired past bio-signals is less than a predetermined number (e.g., 5), by utilizing at least one of the acquired past bio-signals to generate the insufficient number of supplementary bio-signals, and may be generated based on information associated with the acquired past bio-signals and information associated with the supplementary bio-signals.

[0148] In addition, the second encoder module (220) can perform a flattening operation to convert tabular data (30) into one dimension. For example, the second encoder module (220) can convert the tabular data (30) into a one-dimensional vector through a flattening operation, and can generate the second vector after additional processing such as resizing. In one embodiment, the second encoder module (220) can include a 2-1 encoder layer, a 2-2 encoder layer, and a 2-3 encoder layer that are sequentially connected. In addition, the 2-1 encoder layer may include 20 neurons, the 2-2 encoder layer may include 64 neurons, and the 2-3 encoder layer may include 128 neurons, and the second vector including 128 elements may be generated by utilizing these sequential layers. Meanwhile, the second encoder module (220) may generate the second vector by utilizing various neural network architectures such as a fully connected layer, CNN, RNN, and transformer.

[0149] Additionally, the integration module (230) can generate an integration vector based on the first vector and the second vector. For example, the integration module (230) may generate the integrated vector based on the summation of the first vector and the second vector, generate the integrated vector based on the weighted sum of the first vector and the second vector, generate the integrated vector by concatenating the first vector and the second vector, generate the integrated vector by performing element-wise multiplication on the first vector and the second vector, generate the integrated vector by performing average pooling or maximum pooling on the first vector and the second vector, or generate the integrated vector by applying an attention mechanism between the first vector and the second vector. In addition to these methods, the integrated vector may be generated using various methods.

[0150] Here, the first vector and the second vector include different numbers of elements, and the integrated vector may include more elements of the first vector than elements of the second vector. That is, the integrated vector may be generated in a form that places more weight on the bio-signal at the time of diagnosis than on information associated with a plurality of past bio-signals. For example, referring to Fig. 10, the integrated vector may be generated based on 2048 elements of the first vector and 128 elements of the second vector. However, the present invention is not limited thereto, and various embodiments may exist.

[0151] According to one embodiment of the present disclosure, the prediction module (300) may generate disease prediction information based on data related to the biosignal generated by the preprocessing module (200) and data related to the past biosignal generated by the preprocessing module (200). For example, the prediction module (300) may include a fully connected layer module (310) that generates the disease prediction information based on the integrated vector, and may include various alternative neural network architectures in addition to the fully connected layer module (310). Referring to FIG. 10, the fully connected layer module (310) may utilize a first FC layer having 2172 neurons to extract various features of input data and generate intermediate features with high expressiveness. These intermediate features may be transferred to a second FC layer and connected to a single output neuron for final prediction. An activation function such as a sigmoid or softmax used in a classification task may be applied to the second FC layer. However, the present invention is not limited thereto. In one embodiment, the prediction module (300) can generate a prediction value between 0 and 1 that diagnoses heart failure (HFrEF).

[0152]

[0153] Table 1 shows the experimental results of a performance test of a "model for diagnosing heart failure (HFrEF) using electrocardiogram (ECG) signals" implemented according to one embodiment of the present disclosure. In the experiment, training was performed based on binary classification with labels of HFrEF (1) and non-HFrEF (0), and performance was evaluated using a test set separated from the training data. In addition, the model's performance was evaluated based on the Area Under the Receiver Operating Characteristic (AUROC) evaluation metric.

[0154] Referring to Table 1, it was confirmed that, compared to a model utilizing a conventional single-point electrocardiogram (ECG) signal (i.e., a model utilizing only the bio-signal at the time of diagnosis), when disease prediction information is generated using a neural network model according to an embodiment of the present disclosure, performance is improved under all conditions based on the AUROC metric. For example, performance improvements were confirmed in all of the following examples: ① In addition to the biosignal at the time of diagnosis, information associated with 5 past biosignals is further utilized, and EF (ejection fraction), which is past medical information, is included in the associated information (K=5, past EF is utilized); ② In addition to the biosignal at the time of diagnosis, information associated with 5 past biosignals is further utilized, and EF (ejection fraction), which is past medical information, is not included in the associated information (K=5, past EF is not utilized); ③ In addition to the biosignal at the time of diagnosis, information associated with 3 past biosignals is further utilized, and EF (ejection fraction), which is past medical information, is included in the associated information (K=3, past EF is utilized); ④ In addition to the biosignal at the time of diagnosis, information associated with 3 past biosignals is further utilized, and EF (ejection fraction), which is past medical information, is not included in the associated information (K=3, past EF is not utilized). In particular, performance improvements were confirmed compared to single-time-based models in both internal validation and external validation even when past medical information (EF; ejection fraction) history was not utilized.

[0155] Internal validationExternal validationSingle ECG model0.93980.9085Proposed model (K=5)Used past EF0.95560.9425Does not use past EF0.94360.9192Proposed model (K=3)Used past EF0.95700.9392Does not use past EF0.94960.9200

[0156] In the above description, steps S110 to S130 may be further divided into additional steps or combined into fewer steps, depending on the implementation example of the present invention. Furthermore, some steps may be omitted as needed, and the order of steps may be changed.

[0157] According to one embodiment of the present disclosure, a computer-readable medium storing a data structure is disclosed.

[0158] A data structure can refer to the organization, management, and storage of data to enable efficient access and modification. A data structure can refer to the organization of data to solve a specific problem (e.g., data retrieval, data storage, or data modification in the shortest possible time). A data structure can also be defined as the physical or logical relationships between data elements designed to support specific data processing functions. Logical relationships between data elements can include connections between user-defined data elements. Physical relationships between data elements can include actual relationships between data elements physically stored on a computer-readable storage medium (e.g., persistent storage). Specifically, a data structure can include a collection of data, relationships between data, and functions or commands applicable to the data. An effectively designed data structure allows a computing device to perform operations while minimizing the use of its resources. Specifically, a computing device can improve the efficiency of operations, reading, inserting, deleting, comparing, exchanging, and searching through an effectively designed data structure.

[0159] Data structures can be categorized as linear or nonlinear, depending on their form. A linear data structure can be a structure in which only one data item is linked to the next. Linear data structures can include lists, stacks, queues, and deques. A list can refer to a series of data sets with an internal order. Lists can also include linked lists. A linked list is a data structure in which data is linked in a single line, each item having a pointer. In a linked list, a pointer can contain information about the next or previous item. Linked lists can be expressed as singly linked lists, doubly linked lists, or circular linked lists, depending on their form. A stack can be a data listing structure with limited data access. A stack can be a linear data structure in which data operations (e.g., insertion or deletion) can only be performed at one end of the data structure. Data stored in a stack can be a Last-in-First-out (LIFO) data structure. A queue is a data structure with limited access to data. Unlike a stack, it can be a first-in, first-out (FIFO) data structure, with later data being retrieved later. A deck can be a data structure that can process data at both ends.

[0160] A nonlinear data structure can be a structure in which multiple data are connected behind one data. A nonlinear data structure can include a graph data structure. A graph data structure can be defined by vertices and edges, and an edge can include a line connecting two different vertices. A graph data structure can include a tree data structure. A tree data structure can be a data structure in which there is only one path connecting two different vertices among multiple vertices included in the tree. In other words, it can be a data structure that does not form a loop in a graph data structure.

[0161] Throughout this specification, the terms computational model, neural network, network function, and neural network may be used interchangeably. Hereinafter, they are collectively referred to as neural networks. The data structure may include a neural network. And the data structure including the neural network may be stored on a computer-readable medium. The data structure including the neural network may also include preprocessed data for processing by the neural network, data input to the neural network, weights of the neural network, hyperparameters of the neural network, data obtained from the neural network, activation functions associated with each node or layer of the neural network, loss functions for learning the neural network, etc. The data structure including the neural network may include any of the components disclosed above. That is, the data structure including the neural network may be configured to include all or any combination of preprocessed data for processing by the neural network, data input to the neural network, weights of the neural network, hyperparameters of the neural network, data obtained from the neural network, activation functions associated with each node or layer of the neural network, loss functions for learning the neural network, etc. In addition to the aforementioned configurations, a data structure including a neural network may include any other information that determines the characteristics of the neural network. Furthermore, the data structure may include any form of data used or generated in the computational process of the neural network, and is not limited to the aforementioned. The computer-readable medium may include a computer-readable recording medium and / or a computer-readable transmission medium. A neural network may be composed of a set of interconnected computational units, which may generally be referred to as nodes. These nodes may also be referred to as neurons. A neural network is composed of at least one node.

[0162] The data structure may include data input to a neural network. The data structure including the data input to the neural network may be stored on a computer-readable medium. The data input to the neural network may include training data input during the neural network training process and / or input data input to the neural network after training has been completed. The data input to the neural network may include data that has undergone preprocessing and / or data that is the target of preprocessing. Preprocessing may include a data processing process for inputting data to the neural network. Accordingly, the data structure may include data that is the target of preprocessing and data generated by the preprocessing. The above-described data structure is merely an example, and the present disclosure is not limited thereto.

[0163] The data structure may include weights of the neural network. (In this specification, the terms "weight" and "parameter" may be used interchangeably.) The data structure including the weights of the neural network may be stored in a computer-readable medium. The neural network may include a plurality of weights. The weights may be variable and may be varied by a user or an algorithm so that the neural network can perform a desired function. For example, when one or more input nodes are interconnected to one output node by respective links, the output node may determine a data value output from the output node based on values ​​input to the input nodes connected to the output node and weights set for links corresponding to each input node. The above-described data structure is merely an example, and the present disclosure is not limited thereto.

[0164] By way of example and not limitation, the weights may include weights that vary during the neural network training process and / or weights that have completed neural network training. The weights that vary during the neural network training process may include weights at the start of the training cycle and / or weights that vary during the training cycle. The weights that have completed neural network training may include weights that have completed the training cycle. Accordingly, a data structure including the weights of a neural network may include a data structure including weights that vary during the neural network training process and / or weights that have completed neural network training. Therefore, the above-described weights and / or combinations of each weight are included in the data structure including the weights of a neural network. The above-described data structures are merely examples and the present disclosure is not limited thereto.

[0165] A data structure including neural network weights can be stored in a computer-readable storage medium (e.g., memory, hard disk) after going through a serialization process. Serialization can be a process of converting a data structure into a form that can be stored on the same or different computing devices and later reconstructed and used. A computing device can serialize the data structure to transmit and receive data over a network. The serialized data structure including neural network weights can be reconstructed on the same or different computing devices through deserialization. The data structure including neural network weights is not limited to serialization. Furthermore, the data structure including neural network weights can include a data structure that increases computational efficiency while minimizing the use of computing device resources (e.g., a B-Tree, a Trie, an m-way search tree, an AVL tree, a Red-Black Tree in nonlinear data structures). The foregoing is merely an example, and the present disclosure is not limited thereto.

[0166] The data structure may include hyperparameters of a neural network. Furthermore, the data structure including the hyperparameters of the neural network may be stored on a computer-readable medium. The hyperparameters may be variables that can be varied by the user. The hyperparameters may include, for example, a learning rate, a cost function, the number of learning cycle repetitions, weight initialization (e.g., setting a range of weight values ​​to be subject to weight initialization), and the number of hidden units (e.g., the number of hidden layers, the number of nodes in the hidden layer). The above-described data structure is merely an example, and the present disclosure is not limited thereto.

[0167]

[0168] FIG. 11 is a simplified, general schematic diagram of an exemplary computing environment in which embodiments of the present disclosure may be implemented.

[0169] Although the present disclosure has been described above as being generally implemented by a computing device, those skilled in the art will appreciate that the present disclosure may be implemented in combination with computer-executable instructions and / or other program modules that may be executed on one or more computers and / or as a combination of hardware and software.

[0170] Generally, program modules include routines, programs, components, data structures, and the like that perform particular tasks or implement particular abstract data types. Furthermore, those skilled in the art will appreciate that the methods of the present disclosure can be implemented with other computer system configurations, including single-processor or multiprocessor computer systems, minicomputers, mainframe computers, as well as personal computers, handheld computing devices, microprocessor-based or programmable consumer electronics, and the like, each of which may be operatively connected to one or more associated devices.

[0171] The described embodiments of the present disclosure can also be practiced in distributed computing environments, where certain tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote memory storage devices.

[0172] Computers typically include a variety of computer-readable media. Computer-readable media can be any media that can be accessed by a computer, and includes both volatile and nonvolatile media, transitory and non-transitory media, removable and non-removable media. By way of example, and not limitation, computer-readable media can include computer-readable storage media and computer-readable transmission media. Computer-readable storage media includes both volatile and nonvolatile media, transitory and non-transitory media, removable and non-removable media implemented in any method or technology for storing information such as computer-readable instructions, data structures, program modules, or other data. Computer-readable storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital video disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be accessed by a computer and used to store the desired information.

[0173] Computer-readable transmission media typically includes any information delivery media that embodies computer-readable instructions, data structures, program modules, or other data in a modulated data signal, such as a carrier wave or other transport mechanism. The term modulated data signal means a signal that has one or more of its characteristics set or changed so as to encode information in the signal. By way of example, and not limitation, computer-readable transmission media includes wired media, such as a wired network or direct-wired connection, and wireless media, such as acoustic, RF, infrared, or other wireless media. Combinations of any of the above are also intended to be included within the scope of computer-readable transmission media.

[0174] An exemplary environment (1100) implementing various aspects of the present disclosure is illustrated, including a computer (1102) comprising a processing unit (1104), system memory (1106), and a system bus (1108). The system bus (1108) connects system components, including but not limited to the system memory (1106), to the processing unit (1104). The processing unit (1104) may be any of a variety of commercially available processors. Dual processors and other multiprocessor architectures may also be utilized as the processing unit (1104).

[0175] The system bus (1108) may be any of several types of bus structures that may be additionally interconnected to a memory bus, a peripheral bus, and a local bus using any of a variety of commercial bus architectures. The system memory (1106) includes read-only memory (ROM) (1110) and random access memory (RAM) (1112). A basic input / output system (BIOS) is stored in non-volatile memory (1110), such as ROM, EPROM, or EEPROM, and includes basic routines that help transfer information between components within the computer (1102), such as during start-up. The RAM (1112) may also include high-speed RAM, such as static RAM, for caching data.

[0176] The computer (1102) also includes an internal hard disk drive (HDD) (1114) (e.g., EIDE, SATA) - which may also be configured for external use within a suitable chassis (not shown), a magnetic floppy disk drive (FDD) (1116) (e.g., for reading from or writing to a removable diskette (1118)), and an optical disk drive (1120) (e.g., for reading from or writing to a CD-ROM disk (1122) or other high-capacity optical media such as a DVD). The hard disk drive (1114), the magnetic disk drive (1116), and the optical disk drive (1120) may be connected to the system bus (1108) by a hard disk drive interface (1124), a magnetic disk drive interface (1126), and an optical drive interface (1128), respectively. The interface (1124) for implementing an external drive includes at least one or both of Universal Serial Bus (USB) and IEEE 1394 interface technologies.

[0177] These drives and their associated computer-readable media provide non-volatile storage of data, data structures, computer-executable instructions, and the like. In the case of the computer (1102), the drives and media correspond to storing any data in a suitable digital format. While the description of computer-readable media above refers to HDDs, removable magnetic disks, and removable optical media such as CDs or DVDs, those skilled in the art will appreciate that other types of media readable by a computer, such as zip drives, magnetic cassettes, flash memory cards, cartridges, and the like, may also be used in the exemplary operating environment, and that any such media may contain computer-executable instructions for performing the methods of the present disclosure.

[0178] A number of program modules, including an operating system (1130), one or more application programs (1132), other program modules (1134), and program data (1136), may be stored in the drive and RAM (1112). All or portions of the operating system, applications, modules, and / or data may also be cached in RAM (1112). It will be appreciated that the present disclosure may be implemented in various commercially available operating systems or combinations of operating systems.

[0179] A user may enter commands and information into the computer (1102) via one or more wired / wireless input devices, such as a keyboard (1138) and a pointing device such as a mouse (1140). Other input devices (not shown) may include a microphone, an IR remote control, a joystick, a game pad, a stylus pen, a touch screen, and the like. These and other input devices are often connected to the processing unit (1104) via an input device interface (1142) that is connected to the system bus (1108), but may be connected by other interfaces such as a parallel port, an IEEE 1394 serial port, a game port, a USB port, an IR interface, and the like.

[0180] A monitor (1144) or other type of display device is also connected to the system bus (1108) via an interface, such as a video adapter (1146). In addition to the monitor (1144), the computer typically includes other peripheral output devices (not shown), such as speakers, a printer, and so on.

[0181] The computer (1102) may operate in a networked environment using logical connections to one or more remote computers, such as remote computer(s) (1148), via wired and / or wireless communications. The remote computer(s) (1148) may be a workstation, a computing device computer, a router, a personal computer, a portable computer, a microprocessor-based entertainment device, a peer device, or other conventional network node, and generally include many or all of the components described for the computer (1102), although for simplicity, only the memory storage device (1150) is shown. The logical connections shown include wired / wireless connections to a local area network (LAN) (1152) and / or a larger network, such as a wide area network (WAN) (1154). Such LAN and WAN networking environments are common in offices and companies and facilitate enterprise-wide computer networks, such as intranets, all of which may be connected to a worldwide computer network, such as the Internet.

[0182] When used in a LAN networking environment, the computer (1102) is connected to a local network (1152) via a wired and / or wireless communication network interface or adapter (1156). The adapter (1156) may facilitate wired or wireless communications to the LAN (1152), which may also include a wireless access point installed therein for communicating with the wireless adapter (1156). When used in a WAN networking environment, the computer (1102) may include a modem (1158), be connected to a communications computing device on the WAN (1154), or have other means of establishing communications over the WAN (1154), such as via the Internet. The modem (1158), which may be internal or external and wired or wireless, is connected to the system bus (1108) via a serial port interface (1142). In a networked environment, program modules or portions thereof described for the computer (1102) may be stored in a remote memory / storage device (1150). It will be appreciated that the network connections depicted are exemplary and other means of establishing a communications link between the computers may be used.

[0183] The computer (1102) operates to communicate with any wireless device or object that is arranged and operates via wireless communication, such as a printer, a scanner, a desktop and / or portable computer, a portable data assistant (PDA), a communication satellite, any equipment or location associated with a radio-detectable tag, and a telephone. This includes at least Wi-Fi and Bluetooth wireless technologies. Accordingly, the communication may be a predefined structure as in a conventional network, or may simply be an ad hoc communication between at least two devices.

[0184] Wi-Fi (Wireless Fidelity) enables connections to the Internet and other devices without wires. Wi-Fi is a wireless technology that allows devices, such as computers, to send and receive data anywhere within the coverage area of ​​a base station, both indoors and outdoors, similar to cell phones. Wi-Fi networks use wireless technologies called IEEE 802.11 (a, b, g, etc.) to provide secure, reliable, and high-speed wireless connections. Wi-Fi can be used to connect computers to each other, to the Internet, and to wired networks (using IEEE 802.3 or Ethernet). Wi-Fi networks can operate in the unlicensed 2.4 and 5 GHz radio bands, at data rates of, for example, 11 Mbps (802.11a) or 54 Mbps (802.11b), or in products that include both bands (dual-band).

[0185] Those skilled in the art will appreciate that information and signals may be represented using any of a variety of different technologies and techniques. For example, the data, instructions, commands, information, signals, bits, symbols, and chips referenced in the above description may be represented by voltages, currents, electromagnetic waves, magnetic fields or particles, optical fields or particles, or any combination thereof.

[0186] Those skilled in the art will appreciate that the various illustrative logical blocks, modules, processors, means, circuits, and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, various forms of programs or design code (referred to herein, for convenience, as software), or a combination of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. Those skilled in the art may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present disclosure.

[0187] The various embodiments presented herein can be implemented as a method, apparatus, or article of manufacture using standard programming and / or engineering techniques. The term article of manufacture includes a computer program, carrier, or media accessible from any computer-readable storage device. For example, computer-readable storage media include, but are not limited to, magnetic storage devices (e.g., hard disks, floppy disks, magnetic strips, etc.), optical disks (e.g., CDs, DVDs, etc.), smart cards, and flash memory devices (e.g., EEPROMs, cards, sticks, key drives, etc.). Furthermore, various storage media presented herein include one or more devices and / or other machine-readable media for storing information.

[0188] It should be understood that the specific order or hierarchy of steps in the presented processes is merely an example of exemplary approaches. It should be understood that the specific order or hierarchy of steps in the processes may be rearranged within the scope of the present disclosure based on design priorities. The appended method claims provide elements of various steps in a sample order, but are not intended to be limited to the specific order or hierarchy presented.

[0189] The description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other embodiments without departing from the scope of the present disclosure. Therefore, the present disclosure is not intended to be limited to the embodiments disclosed herein, but is to be construed in the broadest scope consistent with the principles and novel features disclosed herein.

[0190] As described above, the relevant contents have been described in the best form for carrying out the invention.

Claims

1. A method for generating disease prediction information by utilizing biosignals, performed by a computing device. A step of acquiring bio-signals at the time of diagnosis; A step of acquiring a past bio-signal of a past point in time prior to the diagnosis point in time or information associated with the past bio-signal; and A step of generating disease prediction information for the time of diagnosis based on the bio-signal at the time of diagnosis, the past bio-signal at the past time, or information associated with the past bio-signal, by utilizing a neural network model. including, method.

2. In paragraph 1, The step of generating disease prediction information for the above diagnosis time is: A step of generating a first vector based on the above biosignal; A step of generating a second vector based on the past biosignal or information associated with the past biosignal; and A step of generating disease prediction information for the diagnosis time point based on the first vector and the second vector by utilizing the neural network model. including, method.

3. In paragraph 2, Information related to the above past bio-signals, Information on the time difference between the above past point in time and the above diagnosis point in time; Medical information from the above past time; or Reference prediction information generated for the past biosignals using a reference neural network model different from the above neural network model. Containing at least one piece of information, method.

4. In paragraph 3, The medical information of the past point in time includes medical measurement information measured within a predetermined time from the time of acquisition of the past bio-signal. method.

5. In paragraph 3, The above reference neural network model is a model that generates a reference prediction value based on a biosignal at a single point in time. method.

6. In paragraph 2, The step of generating the disease prediction information for the diagnosis time point based on the first vector and the second vector by utilizing the neural network model is as follows: A step of generating an integrated vector based on the first vector and the second vector; and A step of generating disease prediction information for the diagnosis time based on the integrated vector by utilizing the above neural network model. including, method.

7. In paragraph 6, The above first vector and the above second vector are vectors generated through different preprocessing methods, The first vector and the second vector contain different numbers of elements, The integrated vector contains more elements of the first vector than elements of the second vector, method.

8. In paragraph 2, The step of acquiring past bio-signals or information related to past bio-signals from a past point in time prior to the above diagnosis point in time is as follows: A step of acquiring a plurality of past bio-signals for a plurality of past points in time prior to the diagnosis point in time or information associated with the plurality of past bio-signals, The step of generating a second vector based on the past biosignal or information associated with the past biosignal is: A step of generating the second vector based on information associated with the plurality of past bio-signals, method.

9. In paragraph 8, Information associated with the above multiple past bio-signals, Each individual piece of information associated with each past biosignal included in the plurality of past biosignals is included, Each of the above individual pieces of information associated with each of the above past bio-signals, Information on the time difference between each of the above past points in time and the above diagnosis point in time; Medical information mapped to each of the above past bio-signals; or Reference prediction information generated for each of the above past biosignals using a reference neural network model different from the above neural network model. Containing at least one piece of information, method.

10. In paragraph 8, The step of generating the second vector based on information associated with the plurality of past bio-signals is as follows: When the number of the plurality of past bio-signals is less than a predetermined number, a step of generating the insufficient number of supplementary bio-signals by utilizing at least one past bio-signal among the plurality of past bio-signals; and A step of generating the second vector based on information associated with the plurality of past bio-signals and the supplementary bio-signals. including, method.

11. In Article 10 The above supplementary bio-signals are generated based on the past bio-signal of the most recent point in time among the plurality of past bio-signals. method.

12. In paragraph 8, Information associated with the above multiple past bio-signals, Includes tabular data generated based on the above multiple past bio-signals, The data in the above table format is: The first type of data related to time differences; A second type of data related to medical information; A third type of data related to disease prediction based on a single point in time; or The fourth type of data related to missing information Contains at least one type of data, method.

13. A structure of a neural network model implemented by a computing device, A preprocessing module for preprocessing a biosignal at the time of diagnosis and preprocessing a past biosignal at a past time before the time of diagnosis or information associated with the past biosignal; and A prediction module that generates disease prediction information based on data related to the biosignal generated by the preprocessing module and data related to the past biosignal generated by the preprocessing module. including, Structure of a neural network model.

14. In paragraph 13, The above preprocessing module, A first encoder module that generates a first vector based on the biosignal at the time of diagnosis; A second encoder module that generates a second vector based on the past biosignal or information associated with the past biosignal; and An integration module that generates an integration vector based on the first vector and the second vector. including, Structure of a neural network model.

15. In paragraph 14, The above prediction module, Generating the disease prediction information based on the above integrated vector, Structure of a neural network model.

16. A computer program stored in a computer-readable storage medium, wherein when the computer program is executed on one or more processors, the computer program causes the one or more processors to perform the following operations for generating disease prediction information using biosignals, wherein the operations are: The act of acquiring biosignals at the time of diagnosis; An operation of acquiring a past bio-signal from a past point in time prior to the above diagnosis time or information associated with the past bio-signal; and An operation of generating disease prediction information for the time of diagnosis based on the bio-signal at the time of diagnosis, the past bio-signal at the past time, or information associated with the past bio-signal, by utilizing a neural network model. including, A computer program stored on a computer-readable storage medium.

17. As a computing device, at least one processor; and memory Including, At least one processor, Obtain vital signs at the time of diagnosis; Obtaining past bio-signals from a past point in time prior to the above diagnosis time or information associated with the past bio-signals; and By utilizing a neural network model, disease prediction information for the time of diagnosis is generated based on the bio-signal at the time of diagnosis, the past bio-signal at the past time, or information associated with the past bio-signal. device.

Images