Health status prediction system utilizing asynchronous electrocardiogram

Abstract

A health state prediction system using asynchronous electrocardiogram, comprising: a single-lead electrocardiogram measurement unit (110) provided with two electrodes, which measures electrocardiogram of two or more electric axes with time difference to obtain asynchronous electrocardiogram data; and a prediction unit (120) which predicts disease and disease degree according to asynchronous electrocardiogram data input from the single-lead electrocardiogram measurement unit (110) and segmented in a specific time unit by a diagnostic algorithm (121) which is a diagnostic algorithm established in advance after learning a plurality of asynchronous segmented standard electrocardiogram data sets which match asynchronous segmented standard lead electrocardiogram and disease and disease degree corresponding to the asynchronous segmented standard lead electrocardiogram, the asynchronous segmented standard lead electrocardiogram being asynchronous segmented standard lead electrocardiogram of different time from synchronous standard lead electrocardiogram measured in the same electric axis and segmented in a specific time unit and accumulated in a medical institution server, the present application generates asynchronous electrocardiogram by segmenting synchronous electrocardiogram of a medical institution and generates a data set for prediction model learning, and can predict disease from asynchronous electrocardiogram data of two or more leads measured by a single-lead electrocardiogram machine.

Description

Technical Field

[0001] This invention discloses a health status prediction system using asynchronous electrocardiograms. The system segments and extracts asynchronous electrocardiograms with two or more leads from various forms of synchronous electrocardiograms accumulated by medical institutions and generates a dataset for predictive model learning. It can predict diseases from asynchronous electrocardiogram data with two or more leads measured by a single-lead electrocardiograph. Background Technology

[0002] As is well known, since the development of the electrocardiogram (ECG), knowledge related to ECG has increased exponentially. Information about the heart's electrical function can be obtained from ECG examinations, and various heart diseases such as arrhythmias, coronary artery disease, and myocardial disease can be further diagnosed.

[0003] Recently, people have been actively researching AI algorithms for electrocardiograms, using AI algorithms to detect heart failure, predict atrial fibrillation in arrhythmic heartbeats, or determine gender.

[0004] As mentioned earlier, it overcomes human limitations and can detect subtle changes in electrocardiogram (ECG) waveforms using AI algorithms, while also improving ECG analysis.

[0005] On the other hand, the electrocardiogram (ECG) used in the medical field is a 12-lead ECG, which is measured by connecting 10 electrodes, including 3 electrodes for the limbs, 6 electrodes for the chest, and 1 electrode for grounding. It can transmit the measured ECG data over long distances.

[0006] However, it would be very inconvenient to use the device in daily life if the chest is exposed and 10 electrodes are connected. Therefore, portable patch-type measuring devices that can perform single-lead electrocardiograms, Samsung smartwatches (Galaxy Watch), Apple smartwatches, etc. are also used.

[0007] A single-lead electrocardiogram measured in this way can be used to diagnose arrhythmias, but its use is limited when diagnosing diseases such as myocardial infarction that require information from various leads of electrocardiogram.

[0008] Therefore, there is an urgent need to develop a technology that can easily measure asynchronous electrocardiograms with two or more leads using a single-lead or six-lead electrocardiograph and predict health status from the measured asynchronous electrocardiogram data with two or more leads. Summary of the Invention

[0009] Technical issues

[0010] The technical challenge to be achieved by this invention is to provide a health status prediction system using asynchronous electrocardiograms (ECGs). This system segments and extracts asynchronous ECGs with two or more leads from various forms of synchronous ECGs accumulated by medical institutions and generates a dataset for learning a prediction model. This system can predict diseases from asynchronous ECG data with two or more leads measured by a single-lead ECG machine.

[0011] Technical solution

[0012] To achieve the aforementioned objective, embodiments of the present invention disclose a health status prediction system utilizing asynchronous electrocardiograms (ECGs), comprising: a single-lead ECG measurement unit equipped with two electrodes, which measures ECGs on two or more electrical axes at time intervals to obtain asynchronous ECG data; and a prediction unit that uses a diagnostic algorithm to predict the presence and severity of illness based on asynchronous ECG data input from the single-lead ECG measurement unit and segmented in a specific time unit. This diagnostic algorithm is a pre-established algorithm that learns from a plurality of asynchronous segmented standard ECG datasets. These datasets match asynchronous segmented standard lead ECGs with the presence and severity of corresponding diseases. The asynchronous segmented standard lead ECGs are asynchronous segmented standard lead ECGs at different times, segmented from synchronous standard lead ECGs measured on the same electrical axis and accumulated on a medical institution server, using the specific time unit.

[0013] This also includes an electrocardiogram (ECG) data generation unit, which identifies the characteristics of the plurality of asynchronous ECG data measured by the single-lead ECG measurement unit as specific lead ECG data, and generates a plurality of other standard lead ECG data that do not belong to the identified specific lead ECG data. The diagnostic algorithm can take the measured asynchronous ECG data and the generated standard lead ECG data as input, or take the generated standard lead ECG data as input, and output and predict whether there is a disease and the degree of disease.

[0014] Furthermore, the ECG data generation unit can generate synchronized ECG data for multiple standard leads.

[0015] Moreover, the diagnostic algorithm can asynchronously extract the synchronously segmented standard lead electrocardiogram to generate the plurality of asynchronously segmented standard electrocardiogram datasets. The synchronously segmented standard lead electrocardiogram is a synchronously segmented standard lead electrocardiogram that is stored in a synchronous form and segmented by a specific time unit.

[0016] Moreover, the diagnostic algorithm can be established in such a way that it takes standard lead electrocardiogram data (with temporal information removed) from synchronously measured standard lead electrocardiogram data as input and outputs the prevalence and severity of the disease corresponding to the standard lead electrocardiogram data for prediction.

[0017] Moreover, the diagnostic algorithm reflects and learns the individual characteristics of the examinee, and the prediction unit can predict whether and how severe the disease is based on the asynchronous electrocardiogram data from the single-lead electrocardiogram measurement unit after reflecting the individual characteristics.

[0018] Furthermore, the individual characteristic information may include gender and age as demographic information, weight and height and degree of obesity as basic health information, past medical history, medication history and family medical history as disease-related information, and examination information such as blood tests, genetic tests and blood pressure and blood oxygen saturation as vital signs and biological signals.

[0019] The effects of the invention

[0020] According to the present invention, asynchronous electrocardiograms with two or more leads can be easily measured using a single-lead electrocardiograph or a 6-lead electrocardiograph. Various forms of synchronous electrocardiograms accumulated by medical institutions can be segmented and asynchronous electrocardiograms with two or more leads can be extracted to generate a dataset for predictive models to learn from.

[0021] Moreover, diseases can be predicted from asynchronous electrocardiogram data with two or more leads measured by a single-lead electrocardiograph or a 6-lead electrocardiograph.

[0022] Furthermore, multi-channel electrocardiogram data can be generated from asynchronous electrocardiogram data with two or more leads, enabling more accurate output of whether or not a patient has a disease and the severity of the disease for prediction. Attached Figure Description

[0023] Figure 1 This is a schematic diagram illustrating the components of a health status prediction system utilizing asynchronous electrocardiograms according to an embodiment of the present invention.

[0024] Figure 2 This is to show the basis. Figure 1 The flowchart shown is a prediction method for a health status prediction system using asynchronous electrocardiograms. Detailed Implementation

[0025] Embodiments of the present invention having the foregoing features are described in more detail below with reference to the accompanying drawings.

[0026] The health status prediction system based on asynchronous electrocardiograms according to an embodiment of the present invention includes: a single-lead electrocardiogram measurement unit 110, equipped with two electrodes, which measures electrocardiograms of two or more electrical axes at time intervals to obtain asynchronous electrocardiogram data; and a prediction unit 120, which uses a diagnostic algorithm 121 to predict the presence and severity of illness based on the asynchronous electrocardiogram data input from the single-lead electrocardiogram measurement unit 110 and segmented in a specific time unit. The diagnostic algorithm 121 is a pre-established diagnostic algorithm that learns from a plurality of asynchronous segmentation standard electrocardiogram datasets, which segment the asynchronous data... The standard lead electrocardiogram (ECG) and the corresponding asynchronous segmented standard lead ECG are matched to determine the presence and severity of diseases. The asynchronous segmented standard lead ECG is an asynchronous segmented standard lead ECG at different times, which is a synchronous standard lead ECG measured on the same electrical axis and accumulated on a medical institution's server, and is divided into asynchronous standard lead ECGs at specific time units. The asynchronous ECG health status prediction system uses the synchronous ECGs of the medical institution to segment and generate asynchronous ECGs, and generates a dataset for the prediction model to learn from. Diseases are predicted from asynchronous ECG data with two or more leads measured by a single-lead ECG machine.

[0027] The following is a detailed description, with reference to the accompanying drawings, of a health status prediction system using asynchronous electrocardiograms with the aforementioned structure.

[0028] First, the single-lead electrocardiogram measurement unit 110 measures electrocardiograms of two or more electrical axes at different time points by means of having two electrodes that are in contact with two points on the subject's body, and then transmits the asynchronous electrocardiogram data to the prediction unit 120 through the local area network.

[0029] For example, after contacting both hands with electrodes, an electrocardiogram (ECG) of lead I, which corresponds to one electrical axis, is measured. After contacting the right hand and left ankle with electrodes in a combination different from the previous method, an ECG of lead II, which corresponds to another electrical axis, is measured. Multiple asynchronous ECGs targeting two or more different electrical axes can be measured separately.

[0030] Furthermore, the single-lead electrocardiogram measurement unit 110 may include a wearable electrocardiogram patch 111, a smartwatch 112, or a 6-lead electrocardiogram rod that can perform contact or non-contact electrocardiogram measurements in daily life and measure asynchronous or synchronous electrocardiograms. After the electrocardiogram data generation unit 130 generates a plurality of electrocardiograms, they are input into a diagnostic algorithm 121 established based on standard lead electrocardiogram data to predict the corresponding health status.

[0031] Here, the single-lead electrocardiogram measurement unit 110 can also measure the subject's continuous electrocardiogram and transmit it to the prediction unit 120, or measure two electrocardiograms at intervals and transmit them to the prediction unit 120.

[0032] Next, the prediction unit 120 predicts the health status from the segmented asynchronous electrocardiogram data of the single-lead electrocardiogram measurement unit 110 using a pre-learned diagnostic algorithm 121. The diagnostic algorithm 121 predicts the presence and severity of disease based on the asynchronous electrocardiogram data input from the single-lead electrocardiogram measurement unit 110 and segmented in a specific time unit, thereby understanding the health status of the examinee. The diagnostic algorithm 121 is a pre-established diagnostic algorithm after learning from a plurality of asynchronous segmented standard electrocardiogram datasets. The plurality of asynchronous segmented standard electrocardiogram datasets match asynchronous segmented standard lead electrocardiograms with the presence and severity of diseases corresponding to the asynchronous segmented standard lead electrocardiograms. The asynchronous segmented standard lead electrocardiograms are asynchronous segmented standard lead electrocardiograms at different times from synchronous standard lead electrocardiograms measured on the same electrical axis and accumulated on the medical institution's server, segmented in the specific time unit.

[0033] Here, it is possible to obtain asynchronous standard lead electrocardiograms by dividing a standard lead electrocardiogram measured by an electrocardiograph in a medical institution into specific time units (e.g., 2.5-second units) and obtaining standard lead electrocardiograms at different times.

[0034] On the other hand, the diagnostic algorithm 121 is established in such a way that it takes the standard lead electrocardiogram data to be segmented by asynchronous measurement of arbitrary electrical axis as input and outputs the disease status and severity corresponding to the asynchronous segmented standard lead electrocardiogram data for prediction. Therefore, it can predict the health status of the examinee from the asynchronous electrocardiogram data transmitted by the single lead electrocardiogram measurement unit 110.

[0035] Alternatively, the diagnostic algorithm 121 can be established in such a way that it takes standard lead ECG data (with timing information removed) from standard lead ECG data measured synchronously for multiple electrical axes as input and outputs the presence or absence and severity of disease corresponding to the standard lead ECG data for prediction. Therefore, it can predict health status from asynchronous ECG data transmitted by the single lead ECG measurement unit 110.

[0036] Alternatively, the diagnostic algorithm 121 can be established in the following manner: converting multiple asynchronous electrocardiogram data input by the single-lead electrocardiogram measurement unit 110 after time difference measurement into synchronous electrocardiogram data, and predicting whether and how severe the disease is from the converted synchronous electrocardiogram data.

[0037] Alternatively, diagnostic algorithm 121 can also asynchronously extract the synchronously segmented standard lead ECG and generate multiple asynchronously segmented standard ECG datasets. The synchronously segmented standard lead ECG is a synchronously segmented standard lead ECG that is a segmented standard lead ECG data stored in synchronous form and divided into specific time units.

[0038] Therefore, the diagnostic algorithm 121 can match the multiple asynchronous electrocardiogram data segments measured and input by the single-lead electrocardiogram measurement unit 110 with time difference to the asynchronously measured and segmented standard lead electrocardiogram data, or the segmented standard lead electrocardiogram data with time information deleted, or the segmented synchronous standard lead electrocardiogram data.

[0039] Furthermore, by leveraging the ECG data generation unit 130, the individual characteristics of the plurality of asynchronous ECG data measured by the single-lead ECG measurement unit 110 are grasped and identified as specific lead ECG data. The remaining plurality of standard lead ECG data that do not belong to the identified specific lead ECG data are generated. The diagnostic algorithm 121 can take the asynchronous ECG data measured by the single-lead ECG measurement unit 110 and the standard lead ECG data generated by the ECG data generation unit 130 as inputs, or take the standard lead ECG data generated by the ECG data generation unit 130 as inputs, and output whether there is a disease and the degree of disease for prediction.

[0040] For example, when generating multiple standard lead ECG data using the ECG data generation unit 130, individual characteristics are grasped through the inherent characteristics of each lead ECG. The asynchronous ECG data input to the single lead ECG measurement unit 110 is matched with the corresponding specific standard lead ECG data to generate other standard lead ECG data that do not match the specific standard lead ECG data, thereby generating multiple new standard lead ECG data.

[0041] Furthermore, the ECG data generation unit 130 can also generate a plurality of synchronized standard lead ECG data. That is, the standard lead ECG data generated by the ECG data generation unit 130 can be asynchronous or synchronous ECG. As mentioned above, when the diagnostic algorithm 121 uses asynchronously measured standard lead ECG data or uses standard lead ECG data with timing information deleted, the ECG data generation unit 130 can generate asynchronous ECG data or generate ECG data without considering synchronization.

[0042] Therefore, when predicting the presence and severity of a disease that requires multi-lead electrocardiogram (ECG) diagnosis, the asynchronous ECG measured by the single-lead ECG measurement unit 110 is used to perform the prediction and diagnosis. Based on multiple ECG information, the disease can be analyzed more accurately, and health status can be measured, diagnosed, examined, and predicted.

[0043] For example, for diseases like arrhythmias that can be diagnosed with a single lead, using a two-lead electrocardiogram and multiple electrocardiograms generated from it to generate electrocardiograms of each heartbeat on various electrical axes can enable more accurate measurement, diagnosis, examination, and prediction of health status. For diseases like myocardial infarction that can be diagnosed with multiple leads, multiple electrocardiograms can be generated to diagnose myocardial infarction.

[0044] Alternatively, diagnostic algorithm 121 could be a model that measures, diagnoses, examines, and predicts health status using only the equivalent of the two leads of ECG data to be used in standard lead ECG data for the purpose of diagnosing diseases, but as mentioned above, it is not limited to two leads and can also use ECG data from further generated leads.

[0045] That is, the electrocardiogram (ECG) data accumulated by medical institutions is standard 12-lead ECG data. The ECG data generation unit 130 generates a standard 12-lead ECG from the two-lead ECG. The generated ECG is then input into an algorithm that measures, diagnoses, examines, and predicts health status based on the standard 12-lead ECG data from the medical institution. This algorithm can use the two-lead ECG to output more accurate prediction results and predict health status more broadly.

[0046] As an example, when the single-lead electrocardiogram measurement unit 110 performs measurements, if some leads or segments of the electrocardiogram data contain a lot of noise or the electrode contact is lost and the measurement is not accurate, the electrocardiogram data generation unit 130 generates an electrocardiogram of the noise-free lead to fill in the discarded electrocardiogram data, thereby enabling more accurate measurement, diagnosis, examination and prediction of health status.

[0047] Furthermore, the health status is monitored after measuring the baseline electrocardiogram of the subject when he is in a healthy state by the single-lead electrocardiogram measurement unit 110 and the prediction unit 120. Then, the electrocardiograms measured and input in real time by the single-lead electrocardiogram measurement unit 110 in daily life are compared with the baseline electrocardiogram to predict whether the electrocardiogram is measured correctly and whether the subject's health status is abnormal. When an error or abnormality is predicted, a warning message is generated by the alarm unit 140. The single-lead electrocardiogram measurement unit 110 in the form of a smartwatch or an additional smart device can emit a buzzer and transmit the warning message.

[0048] Specifically, in the initial stage of use, after storing the baseline ECG from the single-lead ECG measurement unit 110 and generating additional lead ECGs, it can continuously monitor 12-lead ECGs. After wearing the smartwatch 112 on the right hand and touching the abdomen to measure the ECG of lead II to store as the baseline ECG, the user usually wears the smartwatch 112 on the left hand and touches and measures the ECG of lead I with the right hand. At the same time, it can generate multiple ECG leads, including the ECG of lead II, and can use the smartwatch 112 to perform more diverse health status measurements, diagnoses, examinations and predictions.

[0049] Alternatively, before applying the wearable ECG patch 111, the ECG of lead I can be measured with both hands and stored as a baseline ECG. After applying the wearable ECG patch 111, the ECG of lead V can be measured. Based on the continuously measured or monitored ECG of lead V, multiple synchronized ECGs of leads can be generated, enabling more accurate measurement, diagnosis, examination, and prediction of health status.

[0050] Furthermore, the diagnostic algorithm 121 reflects and learns the individual characteristics of the examinee, and the prediction unit 120 can reflect the individual characteristics and predict whether or not the patient has the disease and the degree of the disease from the asynchronous electrocardiogram data originating from the single-lead electrocardiogram measurement unit 110.

[0051] For example, it can reflect individual characteristic information when using standard lead electrocardiogram data during the learning of the applicable diagnostic algorithm 121 to predict whether and how severe the corresponding disease is. Alternatively, it can reflect the individual characteristic information of the subject whose electrocardiogram is measured by the single lead electrocardiogram measurement unit 110 and generate standard lead electrocardiogram data by the electrocardiogram data generation unit 130 to predict whether and how severe the corresponding disease is.

[0052] Here, individual characteristic information may include the subject's gender and age as demographic information, weight and height and degree of obesity as basic health information, past medical history, medication history and family medical history as disease-related information, and examination information such as blood tests, genetic tests and blood pressure and blood oxygen saturation as vital signs and biological signals.

[0053] On the other hand, the prediction unit 120 can take not only the curve-based electrocardiogram as input, but also the numerical electrocardiogram as input. For example, it also includes a data conversion module 122, which converts the asynchronous electrocardiogram data measured by the single-lead electrocardiogram measurement unit 110 into numerical data equivalent to the electrocardiogram through a specific formula and then generates it. The diagnostic algorithm 121 can also take the numerical data equivalent to the asynchronous electrocardiogram data as input to predict the health status.

[0054] Moreover, the diagnostic algorithm 121 can be built using various deep learning models such as convolutional neural networks, LSTM, RNN, and MLP, as well as various machine learning models such as logistic regression, principle-based models, random forests, and support vector machines.

[0055] For example, the diagnostic algorithm 121 can be built using a deep learning model that uses two or more electrocardiogram (ECG) data measured at different times with time differences. It can integrate two or more ECGs into one ECG data and input it, or input two or more ECGs into the deep learning model separately and extract the semantic features of the intermediate measurement, that is, extract the spatial temporal features. Then, based on this, it compares the two ECGs and measures, diagnoses, examines and predicts the health status or future health status at the time of measurement by the single-lead ECG measurement unit 110.

[0056] Here, as a method for mixing electrocardiograms measured at two or more times, after segmenting a single electrocardiogram according to heartbeat, it is possible to match and input heartbeats of the same lead measured at different times, or to compare semantic features or result values ​​extracted after input.

[0057] Alternatively, instead of segmenting ECGs measured at two or more times by heartbeat, the ECG data can be combined directly. This allows for the direct fusion and input of ECG data from two or more times, or the fusion and input of data from different leads. Alternatively, ECGs measured at different times can be input into the deep learning layer of a deep learning model to extract semantic features or output values, then fused and input into the final conclusion. Alternatively, ECGs from different times can be input into the deep learning layer according to different leads, and the features or output values ​​extracted through deep learning can be fused in the backend to output the final conclusion.

[0058] Here, when using ECGs from two or more moments together, they can be input asynchronously, synchronized by heartbeat, or synchronized using deep learning and then fused together for use.

[0059] That is, a deep learning model can be built using standard lead electrocardiogram data. After developing a deep learning model that can measure, diagnose, examine and predict health status by inputting one electrocardiogram at each time, the model can be used to input electrocardiograms measured at two or more time points. The semantic features or final output values ​​of the deep learning model can be combined to predict the result.

[0060] As mentioned earlier, the final output values ​​can be synthesized using various deep learning methods such as convolutional neural networks, LSTM, RNN, and MLP, as well as various machine learning methods such as logistic regression, principle-based models, random forests, and support vector machines.

[0061] It can diagnose and predict diseases of the circulatory system, endocrine and nutritional and metabolic diseases, tumor diseases, mental and behavioral disorders, nervous system diseases, eye and accessory organ diseases, ear and mastoid diseases, respiratory system diseases, digestive system diseases, skin and skin tissue diseases, musculoskeletal system and connective tissue diseases, genitourinary system diseases, pregnancy and childbirth and postpartum diseases, congenital malformations and deformities, and chromosomal abnormalities through the prediction unit 120 as described above.

[0062] In addition, the prediction unit 120 can confirm the damage caused by physical trauma, confirm the prognosis and measure pain, predict the risk of death or worsening caused by trauma, capture or predict complications, and grasp specific conditions that occur before and after birth.

[0063] Furthermore, in the healthcare field, the predictive unit 120 can measure, diagnose, examine, and predict the health status of the examinee. The examinee's health status can be linked to services such as aging, sleep, weight, blood pressure, blood sugar, blood oxygen saturation, metabolism, stress, tension, fear, alcohol consumption, smoking, problem behaviors, lung capacity, exercise volume, pain management, obesity, body mass, body composition, diet, type of exercise, lifestyle recommendations, emergency management, chronic disease management, medication prescriptions, examination recommendations, examination recommendations, care, remote health management, telemedicine, vaccination and post-vaccination management.

[0064] on the other hand, Figure 2 The basis is shown Figure 1 The flowchart shown is for a health status prediction system using asynchronous electrocardiograms. The following explanation is based on this flowchart.

[0065] First, asynchronous electrocardiogram (ECG) data is obtained by measuring two or more electrical axes at different time points using a single-lead ECG measurement unit 110 equipped with two electrodes and contacting two points on the subject's body. The data is then transmitted to the prediction unit 120 via a local area network (S110).

[0066] Subsequently, the asynchronous electrocardiogram data from the single-lead electrocardiogram measurement unit 110 is segmented by the prediction unit 120 using a pre-learned diagnostic algorithm 121 to generate segmented asynchronous electrocardiogram data (S121). This is a step of predicting health status from the segmented asynchronous electrocardiogram data (S122). The diagnostic algorithm 121 predicts the presence and severity of disease based on the segmented asynchronous electrocardiogram data input from the single-lead electrocardiogram measurement unit 110 to determine the health status. The diagnostic algorithm 121 is a pre-established diagnostic algorithm that learns from a plurality of standard electrocardiogram datasets. The plurality of standard electrocardiogram datasets match the segmented standard lead electrocardiograms after the same electrical axis measurement as the electrocardiograms measured by the single-lead electrocardiogram measurement unit 110, with the presence and severity of diseases corresponding to the segmented standard lead electrocardiograms.

[0067] On the other hand, by means of the ECG data generation unit 130, the individual characteristics of the plurality of asynchronous ECG data measured by the single-lead ECG measurement unit 110 are grasped and identified as specific lead ECG data, and the remaining plurality of standard lead ECG data that do not belong to the identified specific lead ECG data are generated (S130). The diagnostic algorithm 121 can take the asynchronous ECG data measured by the single-lead ECG measurement unit 110 and the standard lead ECG data generated by the ECG data generation unit 130 as inputs or take the standard lead ECG data generated by the ECG data generation unit 130 as inputs and output whether there is a disease and the degree of disease in order to make a prediction.

[0068] Subsequently, the alarm unit 140 can operate as follows: by measuring the baseline electrocardiogram (ECG) of the subject when they are in a healthy state using the single-lead ECG measurement unit 110 and the prediction unit 120, and monitoring their health status, the unit compares the ECG measured and input in real time by the single-lead ECG measurement unit 110 in daily life with the baseline ECG to predict whether the ECG was measured correctly and whether the subject's health status is abnormal. If an error or abnormality is predicted, a warning message is generated. The warning message can be emitted by the single-lead ECG measurement unit 110 in the form of a smartwatch or an additional smart device.

[0069] Therefore, by configuring the health status prediction system using asynchronous electrocardiograms as described above, it is possible to easily measure asynchronous electrocardiograms with two or more leads using a single-lead or 6-lead electrocardiograph. Various forms of synchronous electrocardiograms accumulated by medical institutions can be segmented and used to extract asynchronous electrocardiograms with two or more leads to generate a dataset for predictive model learning. Diseases can be predicted from asynchronous electrocardiogram data with two or more leads measured by a single-lead or 6-lead electrocardiograph. Multi-channel electrocardiogram data can be generated from asynchronous electrocardiogram data with two or more leads to output the presence and severity of disease with higher accuracy for prediction.

[0070] The embodiments and structures shown in the accompanying drawings described in this specification are merely one preferred embodiment of the present invention and do not represent the full technical spirit of the present invention. Therefore, it should be understood that various equivalents and modifications may exist at the time of application of this invention.

Claims

1. A health status prediction system utilizing asynchronous electrocardiogram, characterized in that, include: The single-lead electrocardiogram measurement unit is equipped with two electrodes, which measure electrocardiograms of two or more electrical axes at different time intervals to obtain asynchronous electrocardiogram data; as well as The prediction unit uses a diagnostic algorithm to predict the presence and severity of disease based on asynchronous electrocardiogram (ECG) data input from the single-lead ECG measurement unit and segmented in a specific time unit. This diagnostic algorithm is a pre-established algorithm that learns from multiple asynchronous segmented standard ECG datasets. These multiple asynchronous segmented standard ECG datasets match asynchronous segmented standard lead ECGs with the presence and severity of diseases corresponding to the asynchronous segmented standard lead ECGs. The asynchronous segmented standard lead ECGs are asynchronous segmented standard lead ECGs at different times, segmented from synchronous standard lead ECGs measured on the same electrical axis and accumulated on a medical institution server, using the specific time unit.

2. The health status prediction system using asynchronous electrocardiogram according to claim 1, characterized in that, It also includes an electrocardiogram (ECG) data generation unit, which learns the characteristics of the plurality of asynchronous ECG data measured by the single-lead ECG measurement unit and identifies them as ECG data of a specific lead, and generates a plurality of other standard lead ECG data that do not belong to the identified specific lead ECG data. The diagnostic algorithm takes the measured asynchronous electrocardiogram data and the generated standard lead electrocardiogram data as input, or takes the generated standard lead electrocardiogram data as input, and outputs and predicts whether the patient has the disease and the degree of the disease.

3. The health status prediction system using asynchronous electrocardiogram according to claim 2, characterized in that, The electrocardiogram data generation unit generates synchronized multiple standard lead electrocardiogram data.

4. The health status prediction system using asynchronous electrocardiogram according to claim 1, characterized in that, The diagnostic algorithm asynchronously extracts the synchronously segmented standard lead electrocardiograms to generate the plurality of asynchronously segmented standard electrocardiogram datasets. The synchronously segmented standard lead electrocardiograms are generated by segmenting the standard lead electrocardiogram data stored in a synchronous form into synchronously segmented standard lead electrocardiograms in a specific time unit.

5. The health status prediction system using asynchronous electrocardiogram according to claim 1, characterized in that, The diagnostic algorithm is established in the following manner: it takes standard lead electrocardiogram data (with temporal information removed) from synchronously measured standard lead electrocardiogram data as input and outputs the prevalence and severity of the disease corresponding to the standard lead electrocardiogram data for prediction.

6. The health status prediction system using asynchronous electrocardiogram according to claim 1, characterized in that, The diagnostic algorithm reflects and learns the individual characteristics of the examinee. The prediction unit reflects the individual characteristic information and then predicts whether or not the patient has the disease and the severity of the disease from the asynchronous electrocardiogram data originating from the single-lead electrocardiogram measurement unit.

7. The health status prediction system using asynchronous electrocardiogram according to claim 6, characterized in that, The individual characteristic information includes gender and age as demographic information, weight and height and degree of obesity as basic health information, past medical history, medication history and family medical history as disease-related information, and examination information including blood tests, genetic tests and blood pressure and blood oxygen saturation as vital signs and biological signals.

Images