Method of detecting symptom of atrial fibrillation in normal sinus rhythm and device thereof

Abstract

To provide a method of detecting a symptom of atrial fibrillation in normal sinus rhythm of a user on the basis of a mobile ECG, and a device of the method.SOLUTION: According to various embodiments, provided is a method of detecting a symptom of atrial fibrillation in normal sinus rhythm and a device of the method, the method including steps of: acquiring electrocardiogram data of a patient; generating electrocardiogram data that has been preprocessed on the basis of at least a part of a specific region in the electrocardiogram data; and determining an atrial fibrillation occurrence risk of the patient by using the preprocessed electrocardiogram data and an atrial fibrillation determination model that has learned in advance.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present disclosure relates to a method and apparatus for detecting the onset of atrial fibrillation in normal sinus rhythm. [Background technology]

[0002] Because arrhythmias have unclear symptoms and come in a variety of types, they can be confused with other diseases. In particular, if they progress to atrial fibrillation, serious complications such as heart failure and angina can progress rapidly, so regular diagnosis and attention are required.

[0003] Diagnosis of arrhythmias such as those that cause atrial fibrillation is based on electrocardiogram (ECG) measurements, which are generally performed at specialized medical facilities, making it difficult for patients to have their ECG measured during their daily lives.

[0004] In fact, even in patients suspected of having arrhythmia, atrial fibrillation often does not show any symptoms when an ECG is taken, which can result in patients experiencing life-threatening arrhythmias while going about their daily lives without symptoms.

[0005] Therefore, in order to prevent emergency situations from occurring in patients with heart disease, research is needed into technology that can measure electrocardiograms during daily life and technology that can predict the occurrence of arrhythmias and atrial fibrillation using electrocardiogram data. Summary of the Invention [Problem to be solved by the invention]

[0006] The present disclosure can provide a method and apparatus for detecting the onset of atrial fibrillation in a user's normal sinus rhythm based on a mobile ECG.

[0007] The problems to be solved by the present disclosure are not limited to those mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the following description. [Means for solving the problem]

[0008] According to one embodiment, a method for detecting signs of atrial fibrillation in normal sinus rhythm can be provided, comprising the steps of acquiring electrocardiogram data of a patient, generating pre-processed electrocardiogram data based on at least a specific region in the electrocardiogram data, and determining the patient's risk of developing atrial fibrillation using the pre-processed electrocardiogram data and a pre-trained atrial fibrillation judgment model.

[0009] Here, the step of generating the pre-processed electrocardiogram data may include the step of removing at least a portion of a predetermined first region before a specific P peak in the electrocardiogram data and a predetermined second region after a specific T peak.

[0010] Here, the step of generating the pre-processed electrocardiogram data may include a step of dividing the pre-processed electrocardiogram data into input data of a predetermined first time period to generate an input data set.

[0011] Here, the pre-trained atrial fibrillation judgment model can be configured to output a probability of the patient's risk of developing atrial fibrillation in response to input of the input data set.

[0012] Here, the method for detecting signs of atrial fibrillation in normal sinus rhythm may further include generating an analysis report regarding the risk of developing atrial fibrillation.

[0013] Here, the method for detecting signs of atrial fibrillation in normal sinus rhythm may further include a step of transmitting the analysis report to a predetermined device of a medical institution if the patient's risk of developing atrial fibrillation exceeds a predetermined reference value.

[0014] Here, the pre-trained atrial fibrillation judgment model can be trained by the steps of acquiring training electrocardiogram data including first normal sinus rhythm data of a patient with a history of atrial fibrillation and second normal sinus rhythm data of a patient without a history of atrial fibrillation, labeling the first normal sinus rhythm data and the second normal sinus rhythm data with distinguishing markers, and training a pre-generated atrial fibrillation judgment model using the training electrocardiogram data.

[0015] Here, the pre-trained atrial fibrillation judgment model can be configured to determine the patient's risk of developing atrial fibrillation based on at least a portion of the ST segment and QRS complex of the pre-processed electrocardiogram data.

[0016] According to various embodiments, an apparatus for detecting signs of atrial fibrillation in normal sinus rhythm can be provided, which includes an electrocardiogram data processing unit that acquires electrocardiogram data of a patient and generates preprocessed electrocardiogram data based on at least a specific region in the electrocardiogram data, and an atrial fibrillation determination unit that determines the patient's risk of developing atrial fibrillation using the preprocessed electrocardiogram data and a pre-trained atrial fibrillation determination model.

[0017] Here, the electrocardiogram data processing unit can remove at least a part of a predetermined first region before a specific P peak and a predetermined second region after a specific T peak in the electrocardiogram data.

[0018] Here, the electrocardiogram data processing unit can generate an input data set by dividing the preprocessed electrocardiogram data into input data of a preset first time period.

[0019] Here, the pre-trained atrial fibrillation judgment model can be configured to output a probability of the patient's risk of developing atrial fibrillation in response to input of the input data set.

[0020] The diagnostic device may further include a determination result processing unit that generates an analysis report on the atrial fibrillation risk when the atrial fibrillation risk of the patient exceeds a preset reference value.

[0021] Here, the determination result processing unit can transmit the analysis report to a device of a predetermined medical institution.

[0022] Here, the device further includes an atrial fibrillation determination model learning unit that learns an atrial fibrillation determination model, and the atrial fibrillation determination model learning unit acquires training electrocardiogram data including first normal sinus rhythm data of a patient with a history of atrial fibrillation and second normal sinus rhythm data of a patient without a history of atrial fibrillation, labels the first normal sinus rhythm data and the second normal sinus rhythm data with distinguishing markers, and learns an atrial fibrillation determination model that has been generated in advance using the training electrocardiogram data, thereby learning the pre-trained atrial fibrillation determination model.

[0023] Here, the pre-trained atrial fibrillation judgment model can be configured to determine the patient's risk of developing atrial fibrillation based on at least a portion of the ST segment and QRS complex of the pre-processed electrocardiogram data. [Effects of the Invention]

[0024] According to various embodiments, as described above, based on mobile ECG using a small number of electrodes, an environment is provided in which electrocardiograms can be measured to the extent that real-time monitoring is possible during daily life, thereby making it easy to collect electrocardiogram data.

[0025] According to various embodiments, early detection of signs of atrial fibrillation from electrocardiogram data consisting of normal sinus rhythm measured by mobile ECG has the effect of providing an environment in which necessary measures can be taken to prepare for the occurrence of a patient's emergency medical situation.

[0026] According to various embodiments, a method and apparatus for early detection of hidden signs of atrial fibrillation even in normal sinus rhythm using a mobile electrocardiogram is provided, which contributes to the provision of an innovative medical environment, such as improving the management of patients' cardiac disease through cardiac health monitoring and early diagnosis of the risk of arrhythmia that may occur in patients in medical settings. [Brief explanation of the drawings]

[0027] [Figure 1] FIG. 1 is a diagram illustrating the configuration of an apparatus for detecting the onset of atrial fibrillation in normal sinus rhythm according to one embodiment. [Figure 2] FIG. 2 is a diagram schematically illustrating an electrocardiogram waveform included in electrocardiogram data acquired by a device according to an embodiment. [Figure 3] FIG. 3 is a flow chart illustrating the operational flow of an apparatus for detecting the onset of atrial fibrillation in normal sinus rhythm electrocardiogram data according to one embodiment. [Figure 4] FIG. 4 is a flowchart showing a detailed flow of operations performed by an apparatus according to one embodiment to detect the onset of atrial fibrillation in normal sinus rhythm electrocardiogram data. [Figure 5] FIG. 5 is a flow chart illustrating the sequence of operations performed by an apparatus in connection with detecting the onset of atrial fibrillation in normal sinus rhythm electrocardiogram data, according to one embodiment. [Figure 6] FIG. 6 is a flowchart showing the flow of operations for learning an atrial fibrillation determination model in an apparatus according to one embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0028] Preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Advantages and features of the present invention, as well as methods for achieving these, will become apparent from the detailed description of the embodiments in conjunction with the accompanying drawings. However, the present invention should not be limited to the embodiments disclosed below, and may be realized in various different forms. However, these embodiments are provided merely to ensure complete disclosure of the present invention and to fully convey the scope of the invention to those skilled in the art, and the present invention is defined solely by the claims. Hereinafter, the same reference numerals refer to the same components.

[0029] Although terms such as "first," "second," and "second" are used to describe various elements, components, and / or sections, it is understood that these elements, components, and / or sections are not limited by these terms. These terms are used to distinguish one element, component, or section from another. Therefore, it is understood that a first element, first component, or first section referred to below may be a second element, second component, or second section within the technical spirit of the present invention.

[0030] The terms used herein are for the purpose of describing embodiments only and are not intended to limit the present invention. The singular terms "a," "an," and "the" as used herein include the plural terms unless otherwise specified. The terms "comprises" and / or "made of," as used herein, do not exclude the presence or addition of one or more other components, steps, operations, and / or elements to which a reference is made.

[0031] Unless otherwise defined, all terms (including technical and scientific terms) used herein are used in the sense that they can be commonly understood by a person having ordinary skill in the art to which the present invention belongs. Furthermore, terms defined in common dictionaries should not be interpreted ideally or excessively unless otherwise defined.

[0032] Hereinafter, preferred embodiments of the present disclosure will be described with reference to the drawings. The drawings attached to this specification, together with the detailed description of the invention, serve to further understand the technical concept of the present invention, and therefore the present disclosure should not be interpreted as being limited to only the matters shown in the drawings.

[0033] This disclosure describes a method and apparatus for detecting the onset of atrial fibrillation, and more particularly, a method and apparatus for detecting the onset of atrial fibrillation in normal sinus rhythm using mobile measured electrocardiogram data.

[0034] FIG. 1 is a diagram illustrating the configuration of an apparatus for detecting the onset of atrial fibrillation in normal sinus rhythm according to one embodiment.

[0035] Referring to FIG. 1, a device 100 (hereinafter referred to as "device 100") for detecting signs of atrial fibrillation in normal sinus rhythm includes a processing unit 110, a memory unit 120, and a communication unit 140, and may further include a sensor unit 130.

[0036] The processing unit 110 includes at least one processor, and can process various data for the operation of the device 100 using at least one program (application, tool, plug-in, software, etc.).

[0037] For this purpose, the processing unit 110 can be divided based on the range of functions or operations. For example, the processing unit 110 can be configured to include an electrocardiogram data processing unit 111 that acquires and / or preprocesses electrocardiogram data, an atrial fibrillation determination unit 113 that determines signs of atrial fibrillation from the electrocardiogram data using a pre-trained atrial fibrillation determination model, and a determination result processing unit 115 that processes the atrial fibrillation determination result.

[0038] Furthermore, the processing unit 110 can be configured to further include an atrial fibrillation determination model learning unit 117 that learns an atrial fibrillation determination model.

[0039] The storage unit 120 can store various data to be processed by at least one component (e.g., the processing unit 110 or the communication unit) of the device 100. The data can include, for example, a program for processing control instructions, data processed by the program, or input data and output data related thereto.

[0040] More specifically, the storage unit 120 may include artificial intelligence algorithms based at least in part on artificial neural network algorithms, blockchain algorithms, deep learning algorithms, regression analysis algorithms, and related mechanisms, operators, language models, and big data for processing control instructions.

[0041] According to one embodiment, the storage unit 120 may include an atrial fibrillation determination model 123 configured to determine the onset of atrial fibrillation from input electrocardiogram data of a patient.

[0042] The atrial fibrillation determination model 123 can be configured to include at least some of the algorithms of ResNet, a recurrent neural network (RNN), a long short-term memory recurrent neural network (LSTM), and a convolutional neural network (CNN).

[0043] In addition, the atrial fibrillation determination model 123 can be configured to include at least some of the optimization algorithms: Adagrad, Root Mean Square Propagation (RMSprop), Adaptive Moment Estimation (Adam), Adadelta, Nadam, AMSGrad, Nesterov Accelerated Gradient Descent (NAG), and Nesterov Momentum.

[0044] Furthermore, the storage unit 120 can be configured to further include training electrocardiogram data 121 for training the atrial fibrillation determination model.

[0045] The training electrocardiogram data 121 may include a plurality of electrocardiogram data of a plurality of patients (those who have experienced atrial fibrillation and / or those who have not experienced atrial fibrillation). The training electrocardiogram data 121 may be divided into at least some categories of a training set, a validation set, and a data set for training an atrial fibrillation determination model.

[0046] The sensor unit 130 may include at least one electrocardiogram sensor for determining an electrocardiogram of the patient.

[0047] More specifically, the electrocardiogram sensor may include at least one electrode that contacts the patient's skin and transmits an electrical signal related to the electrocardiogram. In this case, the sensor unit 130 may include at least some of the functions of a mobile electrocardiogram sensor that includes one electrode (1-lead).

[0048] However, the present invention is not limited to this, and the sensor unit 130 may also be configured to include at least some of the functions of a mobile electrocardiogram sensor including multiple electrodes.

[0049] Referring to FIG. 1, the sensor unit 130 is shown as a component of the device 100, but is not limited to this and may be configured to be detachable from the device 100 or may be configured independently outside the device 100.

[0050] When the sensor unit 130 is configured independently of the device 100, the communication unit 140 can receive electrocardiogram data measured from a specific patient from the sensor unit 130.

[0051] In this regard, FIG. 2 is a diagram schematically illustrating an electrocardiogram waveform included in electrocardiogram data acquired by an apparatus according to one embodiment.

[0052] Referring to FIG. 2, an electrocardiogram waveform includes at least a portion of a P wave, a QRS complex, and a T wave, and can be divided into a plurality of intervals and a plurality of segments based on the P wave, a QRS complex, and a T wave.

[0053] More specifically, the electrocardiogram waveform can be divided into at least some intervals and segments of a PR interval, a QT interval, an ST-T interval, a TR interval, a PR segment, and an ST segment.

[0054] Additionally, the sensor unit 130 may be configured to include at least one sensor for measuring a physiological condition of the patient.

[0055] For example, the sensor unit 130 may be configured to include at least one sensor selected from the group consisting of a temperature measurement sensor for measuring the body temperature of the patient and a pulse measurement sensor for measuring the pulse of the patient.

[0056] As described above, the sensor unit 130 has been described as including a temperature measurement sensor and / or a pulse measurement sensor, but is not limited thereto and may be configured to be connected to a temperature measurement sensor and / or a pulse measurement sensor external to the device 100.

[0057] The communication unit 140 may support the establishment of wired communication channels within the device 100 and / or between the device 100 and at least one other device (e.g., a user device or a server), the establishment of wireless communication channels, and the execution of communications via the established communication channels.

[0058] Although not shown in FIG. 1, the device 100 may further include at least one input / output unit.

[0059] The input / output unit may include or be connected to at least a portion of an input unit (not shown) for inputting data such as a keyboard, mouse, or touchpad, and an output unit (not shown) for outputting data such as a display unit (e.g., a display), speaker, or drive unit.

[0060] According to various embodiments of the present invention, device 100 or user device may include at least some of the functionality of a range of information and communication devices, including mobile communication terminals, multimedia terminals, wireline terminals, fixed terminals, and Internet Protocol (IP) terminals.

[0061] The device 100 can be configured to include at least some of the functionality of a workstation or a large-capacity database as a device for processing control instructions, or to be connected via communication.

[0062] Hereinafter, a method in which the device 100 acquires electrocardiogram data measured from a patient and determines signs of atrial fibrillation from normal sinus rhythm in the electrocardiogram data will be described in detail with reference to FIGS.

[0063] In this regard, Fig. 3 is a flowchart showing the operation flow of an apparatus according to one embodiment for detecting signs of atrial fibrillation in normal sinus rhythm of electrocardiogram data. Fig. 4 is a flowchart showing a detailed operation flow of an apparatus according to one embodiment for detecting signs of atrial fibrillation in normal sinus rhythm of electrocardiogram data. Fig. 5 is a flowchart showing the operation flow performed by an apparatus according to one embodiment in association with detecting signs of atrial fibrillation in normal sinus rhythm of electrocardiogram data. And Fig. 6 is a flowchart showing the operation flow of training an atrial fibrillation determination model in an apparatus according to one embodiment.

[0064] First, referring to FIG. 3, in step 301, the electrocardiogram data processor 111 can acquire electrocardiogram data of a patient.

[0065] More specifically, when the device 100 is configured to include the sensor unit 130, the electrocardiogram data processing unit 111 can acquire the electrocardiogram data of the patient measured by the sensor unit 130.

[0066] According to various embodiments, when the sensor unit 130 is configured independently of the device 100, the electrocardiogram data processing unit 111 can acquire the patient's electrocardiogram data measured by the sensor unit 130 via the communication unit 140.

[0067] In step 303, the electrocardiogram data processor 111 can generate pre-processed electrocardiogram data based on at least some specific regions in the electrocardiogram data.

[0068] According to one embodiment, as shown in FIG. 4 , the electrocardiogram data processing unit 111 can remove (401) at least a portion of a predetermined first region before a specific P peak and a predetermined second region after a specific T peak in the electrocardiogram data, and / or generate (403) an input data set by dividing the pre-processed electrocardiogram data into input data of a predetermined first time.

[0069] To explain step 401 more specifically, the electrocardiogram data processing unit 111 can remove noise regions from the electrocardiogram data acquired from the patient to generate pre-processed electrocardiogram data.

[0070] Here, the noise in the electrocardiogram data may include at least a portion of the noise measured during electrocardiogram measurement, such as noise generated at the moment the sensor unit starts or ends electrocardiogram measurement of the patient, and noise generated at the moment electrodes are attached to or detached from the patient.

[0071] For example, the electrocardiogram data processing unit 111 can remove the area before the first identified P-wave peak (p-peak) and / or the area after the last identified T-wave peak (T-peak) in the electrocardiogram data acquired from the patient.

[0072] To explain step 403 more specifically, the electrocardiogram data processing unit 111 can divide the electrocardiogram data from which at least some regions have been removed into predetermined time units to generate pre-processed electrocardiogram data.

[0073] For example, the electrocardiogram data processing unit 111 can divide the electrocardiogram data into preset time units of 5 to 15 seconds, more preferably into 10-second units.

[0074] However, without being limited thereto, the electrocardiogram data processing unit 111 can divide the electrocardiogram data so as to include a specified number of specific signals. For example, the electrocardiogram data processing unit 111 can divide the electrocardiogram data so that the divided electrocardiogram data (preprocessed electrocardiogram data) includes a preset number of 10 to 18 RR peaks and / or a preset number of 5 to 10 RR intervals.

[0075] FIG. 4 shows that step 403 is executed after step 401 is executed, but this is not limiting, and the electrocardiogram data processing unit 111 can execute step 401 after executing step 403.

[0076] Furthermore, in performing the operation of step 303, the electrocardiogram data processing unit 111 may execute only step 401 or only step 403.

[0077] Although not shown in FIG. 4, the electrocardiogram data processing unit 111 can perform operations such as removing noise from the electrocardiogram data using a filter and / or stabilizing the electrocardiogram baseline.

[0078] More specifically, the electrocardiogram data processing unit 111 can remove noise from the electrocardiogram data using a filter that acts on a specific frequency (for example, a Butterworth filter).

[0079] Here, the Butterworth filter can be configured to act on at least a part of the frequencies between 0 and 1 Hz, for example, 0.5 Hz, but is not limited to this, and the frequency at which the Butterworth filter operates can be changed by setting.

[0080] In addition, the electrocardiogram data processing unit 111 can stabilize the reference line of the electrocardiogram data by estimating a reference line based on a moving average in the electrocardiogram data and subtracting the estimated reference line from the electrocardiogram data (or the electrocardiogram waveform of the electrocardiogram data).

[0081] Furthermore, the electrocardiogram data processing unit 111 can scale the electrocardiogram data so that the distribution of input values ​​to the atrial fibrillation determination model 123 becomes uniform. For example, the electrocardiogram data processing unit 111 can scale the electrocardiogram data so that the values ​​fall within a preset range (for example, from -1 to 1). However, the present invention is not limited to this, and the scaling range of the electrocardiogram data can be changed by settings.

[0082] In step 305, the atrial fibrillation determination unit 113 can determine the patient's risk of developing atrial fibrillation using the pre-processed electrocardiogram data and the pre-trained atrial fibrillation determination model.

[0083] More specifically, the atrial fibrillation determination unit 113 can input the preprocessed electrocardiogram data to a pre-trained atrial fibrillation determination model 123 and acquire the output of the atrial fibrillation determination model 123 .

[0084] Here, the atrial fibrillation determination unit 113 can acquire the output of the atrial fibrillation determination model 123 corresponding to the input pre-processed electrocardiogram data as the risk of atrial fibrillation occurring for the patient.

[0085] According to one embodiment, the atrial fibrillation judgment model 123 can be configured to output the risk of atrial fibrillation occurring as a probability value (%) when a plurality of preprocessed electrocardiogram data configured in a first time unit is input.

[0086] However, without being limited to this, the atrial fibrillation judgment model 123 can be configured to output the risk of atrial fibrillation occurring in various types of values, such as outputting the risk of atrial fibrillation occurring as a decimal value with a specific value (e.g., 1) as the maximum value.

[0087] After executing step 305, the atrial fibrillation determination unit 113 can end the embodiment of FIG.

[0088] Also, although not shown in Figures 3 and 4, as shown in Figure 5, the atrial fibrillation judgment unit 113 can generate (501) an analysis report on the risk of atrial fibrillation occurring output from the atrial fibrillation judgment model 123.

[0089] More specifically, the atrial fibrillation determining unit 113 can generate an analysis report including at least some of the patient's atrial fibrillation risk, atrial fibrillation history information, and personal information.

[0090] For this purpose, the memory unit 120 is in a state in which at least some of the patient's atrial fibrillation history information and personal information is stored, or is capable of receiving at least some of the patient's atrial fibrillation history information and personal information from a pre-set server.

[0091] Here, the atrial fibrillation history information may include at least some of the following information: whether or not there is a history of atrial fibrillation, the time of occurrence of atrial fibrillation (e.g., year, month, day, etc.) if there is a history of atrial fibrillation, and information about the patient's atrial fibrillation at the medical hospital, etc. More specifically, the atrial fibrillation history information may include information about the most recent time when atrial fibrillation occurred.

[0092] Here, the personal information may include at least some of various information about the patient, such as the patient's name, age, date of birth, address, height, weight, and contact information of the guardian.

[0093] Thereafter, if the output result of the atrial fibrillation judgment model 123 indicates that the patient's risk of developing atrial fibrillation exceeds a predetermined reference value, the atrial fibrillation judgment unit 113 can transmit (503) the analysis report to a predetermined device at a medical institution.

[0094] Here, the preset medical institution device may include a device of a medical hospital preset for the patient, but is not limited thereto, and the atrial fibrillation determination unit 113 may transmit the generated analysis report to at least some of various preset devices, such as a guardian's device, a device related to emergency rescue services, etc.

[0095] Here, the emergency rescue service may include at least some of the services that allow emergency communication, such as emergency medical services and police services.

[0096] According to one embodiment, the preset reference value for the patient's risk of developing atrial fibrillation may be set to one value or to two or more different values.

[0097] For example, when the preset reference value is set to one value, if the output result of the atrial fibrillation determination model 123, that is, the patient's risk of developing atrial fibrillation, exceeds the reference value, the atrial fibrillation determination unit 113 can transmit an analysis report to a first device (or a first device list) (e.g., the guardian's device and the emergency rescue service's device) that is preset to include at least one of a preset medical institution's device, a guardian's device, and a device related to an emergency rescue service.

[0098] In another example, the preset values ​​may be set to a first reference value and a second reference value higher than the first reference value.

[0099] At this time, if the patient's risk of developing atrial fibrillation exceeds the first reference value and is equal to or less than the second reference value, the atrial fibrillation determination unit 113 can send the analysis report to a second device (or a second device list) (e.g., the guardian's device) that is preset to include at least one of a device of a medical institution, a device of the guardian, and a device related to an emergency rescue service.

[0100] On the other hand, if the patient's risk of developing atrial fibrillation exceeds the second reference value, the atrial fibrillation determination unit 113 can send the analysis report to a third device (or a third device list) that is pre-set to include at least one of a pre-set medical institution device, a guardian device, and a device related to emergency rescue services (e.g., a guardian device, an emergency medical services device, and a police device).

[0101] At least some of the operations 501 and 503 (501 and / or 503) described in Figure 5 may be processed after execution of step 305 in Figure 3. Also, at least some of the operations described in Figure 5 may be executed based on a request from a user (e.g., a patient and / or a designated healthcare professional).

[0102] The atrial fibrillation determination model 123 that outputs the risk of atrial fibrillation occurring in a patient in response to input of preprocessed electrocardiogram data as described above may be in a trained state as shown in FIG.

[0103] 6 is processed by the electrocardiogram data processing unit 111 and / or the atrial fibrillation determination model learning unit 117 (hereinafter referred to as "learning unit 117"). In the following description, the operation in FIG. 6 will be described as being processed by the learning unit 117.

[0104] In step 601, the learning unit 117 can acquire learning electrocardiogram data including first normal sinus rhythm data (first electrocardiogram data) of a patient with a history of atrial fibrillation (past data) and second normal sinus rhythm data (second electrocardiogram data) of a patient without a history of atrial fibrillation.

[0105] In step 603, the learning unit 117 may label at least a portion of the training electrocardiogram data with a specific marker. For example, the learning unit 117 may label the first normal sinus rhythm data and the second normal sinus rhythm data with a distinguishing marker.

[0106] More specifically, the learning unit 117 can set a label value of 1 for the first normal sinus rhythm data and a label value of 0 for the second normal sinus rhythm data.

[0107] However, the present invention is not limited to this, and the label value for the first normal sinus rhythm data and / or the label value for the second normal sinus rhythm data can be changed by settings.

[0108] Furthermore, the learning unit 117 can generate pre-processed training electrocardiogram data by performing at least some of the operations of noise removal, baseline stabilization, and scaling on at least some of the training electrocardiogram data.

[0109] Here, the pre-processing of the learning electrocardiogram data can be performed based on at least a part of the pre-processing operation performed in step 303 .

[0110] In step 605, the learning unit 117 can learn an atrial fibrillation determination model generated in advance using the training electrocardiogram data. More specifically, the learning unit 117 can learn an atrial fibrillation determination model generated in advance using training electrocardiogram data that has been subjected to labeling and / or preprocessing.

[0111] Therefore, the atrial fibrillation judgment model 123 may be the result of (or generated by) executing learning of a previously generated atrial fibrillation judgment model.

[0112] More specifically, the pre-generated atrial fibrillation judgment model may be an artificial intelligence model generated to output the risk of atrial fibrillation occurring based on input electrocardiogram data (and / or pre-processed electrocardiogram data). Based on this, the atrial fibrillation judgment model 123 may be a result of the pre-generated atrial fibrillation judgment model being learned based on at least some of the operations of FIG. 6.

[0113] The pre-generated atrial fibrillation judgment model can be configured to determine the risk of atrial fibrillation occurring based on at least a portion of a particular segment or complex of electrocardiogram data (and / or pre-processed electrocardiogram data) during the learning process, or to assign weight to determining the risk of atrial fibrillation occurring.

[0114] For example, the atrial fibrillation decision model may determine a patient's risk of developing atrial fibrillation based on values ​​(and / or changes in values) of at least a portion of the ST segment and QRS complex of the electrocardiogram data.

[0115] The atrial fibrillation determination model may also be configured to weight the determination of a patient's risk of developing atrial fibrillation based on values ​​(and / or changes in values) of at least some of the ST segment and QRS complex of the electrocardiogram data.

[0116] As described above, the atrial fibrillation judgment model can be trained using electrocardiogram data (or preprocessed electrocardiogram data) consisting of normal sinus rhythm as training electrocardiogram data, and can be trained to determine the risk of atrial fibrillation occurring from electrocardiogram data having normal sinus rhythm.

[0117] Based on this, the atrial fibrillation judgment model 123 can also determine the risk of atrial fibrillation occurring based on at least some values ​​(and / or changes in values) of the ST segment and QRS complex of the electrocardiogram data, or can give weight to determining the risk of atrial fibrillation occurring.

[0118] According to various embodiments, as described above, based on mobile ECG using a small number of electrodes, an environment is provided in which electrocardiograms can be measured to the extent that real-time monitoring is possible during daily life, thereby making it easy to collect electrocardiogram data.

[0119] According to various embodiments, early detection of signs of atrial fibrillation from electrocardiogram data consisting of normal sinus rhythm measured by mobile ECG has the effect of providing an environment in which necessary measures can be taken to prepare for the occurrence of a patient's emergency medical situation.

[0120] According to various embodiments, a method and apparatus for early detection of hidden signs of atrial fibrillation even in a state of normal sinus rhythm using a mobile electrocardiogram are provided, which contributes to the provision of an innovative medical environment, such as improving patient cardiac disease management through cardiac health monitoring and early diagnosis of arrhythmia risks that may occur in patients in medical settings.

[0121] As described above, although the embodiments have been described using limited drawings, a person having ordinary skill in the art can apply various technical modifications and variations based on the various embodiments.

[0122] For example, the techniques described may be performed in a different order than described, and / or the components of the described systems, structures, devices, circuits, etc. may be combined or combined in a different manner than described, or may be substituted or replaced by other components or equivalents, and still achieve suitable results.

[0123] In particular, when a description is given with reference to a flowchart, a plurality of steps are configured and described as being executed sequentially in a specified order, but the order is not necessarily limited to the order described.

[0124] In other words, embodiments may be applied in which at least some of the steps shown in the flowcharts are modified or deleted, or at least one step is added, or one or more steps are executed in parallel. In other words, the steps are not necessarily limited to operating in chronological order, and this is also included in embodiments of the present disclosure.

[0125] Accordingly, other implementations, embodiments, and equivalents of the claims are intended to be within the scope of the following claims. [Explanation of symbols]

[0126] 100: Equipment 110: Processing unit 111: Electrocardiogram data processing unit 113: Atrial fibrillation detection unit 115: Judgment result processing unit 117: Atrial fibrillation judgment model learning unit 120: Storage section 121: Electrocardiogram data for learning 123: Atrial fibrillation judgment model 130: Sensor unit 140: Communications Department

Claims

1. acquiring electrocardiogram data of the patient; generating pre-processed electrocardiogram data based on at least a portion of the specific region in the electrocardiogram data; and determining a patient's risk of developing atrial fibrillation using the pre-processed electrocardiogram data and a pre-trained atrial fibrillation judgment model. A method for detecting the onset of atrial fibrillation in normal sinus rhythm.

2. generating the pre-processed electrocardiogram data includes: removing at least a portion of a first predetermined region before a particular P-peak and a second predetermined region after a particular T-peak from the electrocardiogram data; 2. The method of detecting onset of atrial fibrillation in normal sinus rhythm according to claim 1.

3. generating the pre-processed electrocardiogram data includes: dividing the pre-processed electrocardiogram data into input data of a predetermined first time period to generate an input data set; 2. The method of detecting onset of atrial fibrillation in normal sinus rhythm according to claim 1.

4. The pre-trained atrial fibrillation determination model and configured to output a probability of the patient's risk of developing atrial fibrillation in response to input of the input data set.

4. The method for detecting onset of atrial fibrillation in normal sinus rhythm according to claim 3.

5. generating an analysis report regarding the risk of atrial fibrillation occurring; 2. The method of detecting onset of atrial fibrillation in normal sinus rhythm according to claim 1.

6. If the patient's risk of developing atrial fibrillation exceeds a predetermined reference value, the analysis report is sent to a predetermined device of a medical institution.

6. The method for detecting onset of atrial fibrillation in normal sinus rhythm according to claim 5.

7. The pre-trained atrial fibrillation determination model acquiring training electrocardiogram data including first normal sinus rhythm data of a patient with a history of atrial fibrillation and second normal sinus rhythm data of a patient without a history of atrial fibrillation; labeling the first normal sinus rhythm data and the second normal sinus rhythm data with a marker; and a step of learning an atrial fibrillation determination model generated in advance using the learning electrocardiogram data.

2. The method of detecting onset of atrial fibrillation in normal sinus rhythm according to claim 1.

8. The pre-trained atrial fibrillation determination model determining a risk of the patient developing atrial fibrillation based at least in part on an ST segment and a QRS complex of the preprocessed electrocardiogram data.

2. The method of detecting onset of atrial fibrillation in normal sinus rhythm according to claim 1.

9. an electrocardiogram data processing unit that acquires electrocardiogram data of a patient and generates pre-processed electrocardiogram data based on at least a specific region of the electrocardiogram data; an atrial fibrillation determination unit that determines a patient's risk of developing atrial fibrillation using the preprocessed electrocardiogram data and a pre-trained atrial fibrillation determination model; A device that detects signs of atrial fibrillation in normal sinus rhythm.

10. The electrocardiogram data processing unit removing at least a portion of a first predetermined region before a specific P-peak and a second predetermined region after a specific T-peak in the electrocardiogram data; 10. An apparatus for detecting the onset of atrial fibrillation in normal sinus rhythm as claimed in claim 9.

11. The electrocardiogram data processing unit Dividing the preprocessed electrocardiogram data into input data of a predetermined first time period to generate an input data set; 10. An apparatus for detecting the onset of atrial fibrillation in normal sinus rhythm as claimed in claim 9.

12. The pre-trained atrial fibrillation determination model and configured to output a probability of the patient's risk of developing atrial fibrillation in response to input of the input data set.

12. An apparatus for detecting the onset of atrial fibrillation in normal sinus rhythm as claimed in claim 11.

13. and a determination result processing unit that generates an analysis report on the atrial fibrillation risk of the patient when the atrial fibrillation risk exceeds a predetermined reference value.

10. An apparatus for detecting the onset of atrial fibrillation in normal sinus rhythm as claimed in claim 9.

14. The determination result processing unit transmitting the analysis report to a predetermined device of a medical institution; 14. An apparatus for detecting the onset of atrial fibrillation in normal sinus rhythm as claimed in claim 13.

15. further including an atrial fibrillation determination model learning unit that learns an atrial fibrillation determination model; The atrial fibrillation determination model learning unit acquiring training electrocardiogram data including first normal sinus rhythm data of a patient with a history of atrial fibrillation and second normal sinus rhythm data of a patient without a history of atrial fibrillation; labeling the first normal sinus rhythm data and the second normal sinus rhythm data with a distinguishing marker; a pre-trained atrial fibrillation determination model is trained by using the training electrocardiogram data to train the pre-trained atrial fibrillation determination model.

10. An apparatus for detecting the onset of atrial fibrillation in normal sinus rhythm as claimed in claim 9.

16. The pre-trained atrial fibrillation determination model determining a risk of the patient developing atrial fibrillation based at least in part on an ST segment and a QRS complex of the preprocessed electrocardiogram data.

10. An apparatus for detecting the onset of atrial fibrillation in normal sinus rhythm as claimed in claim 9.

Images