Heart condition detection sensor device and system for providing complex life support solution using same

Patent metadata

Field	Value
Publication number	US-2023282352-A1
Application number	US-202118008266-A
Country	US
Kind code	A1
Filing date	Jun 11, 2021
Priority date	Jun 12, 2020
Publication date	Sep 7, 2023
Grant date	—

Abstract

The present invention relates to a technical idea of monitoring a heart condition by analyzing biosignals measured using a sensor device for detecting a heart condition. In a method of monitoring a heart condition according to one embodiment of the present invention, electrocardiogram signals are measured from a user, feature information is extracted by performing machine learning of the time domain of the measured electrocardiogram signals, a plurality of cardiac abnormality type models are determined by performing machine learning of the extracted feature information, classification accuracy for the determined cardiac abnormality type models is calculated, and a cardiovascular disease of the user is determined using the determined cardiac abnormality type models and public cardiovascular disease data based on the calculated accuracy. That is, the present invention relates to a technique for assisting medical diagnosis.

First claim

1 . A server that is linked to a sensor device for detecting a heart condition and provides user condition information, wherein the server comprises a monitoring information collector for collecting monitoring information comprising biosignals comprising electrocardiogram signals measured from a user;
a signal extractor for extracting the electrocardiogram signals comprised in the collected monitoring information; an artificial intelligence processor for extracting morphological information as feature information by converting the extracted electrocardiogram signals into a time-standardized image based on a pre-stored artificial intelligence machine learning algorithm, determining a plurality of cardiac abnormality type models using the extracted feature information, calculating classification accuracy for the determined cardiac abnormality type models, and determining a cardiovascular disease of the user using the determined cardiac abnormality type models and public cardiovascular disease data based on the calculated accuracy; and a controller for controlling to provide the determined cardiovascular disease to a user terminal.

2 . The server according to claim 1 , wherein the artificial intelligence processor generates a normalized signal based on a time domain of the extracted electrocardiogram signals, converts the generated normalized signal into the time-standardized image, generates a compressed signal by applying the pre-stored artificial intelligence machine learning algorithm-based weight to the converted image, generates a reconstructed signal from the compressed signal using the applied weight, and extracts morphological information of the electrocardiogram signals as the feature information by performing machine learning of the weight so that a difference between the generated normalized signal and the generated reconstructed signal falls within a preset threshold range.

3 . The server according to claim 1 , wherein the artificial intelligence processor performs machine learning of the feature information to determine the cardiac abnormality type models as at least one model of a tachycardia model, a bradycardia model, an atrial fibrillation model, a left bundle branch block model, a right bundle branch block model, a premature atrial contraction model, a premature ventricular contraction model, a cardiac arrest model, and a normal heart condition model.

4 . The server according to claim 3 , wherein the artificial intelligence processor uses an open data set, at least one model of the tachycardia model, the bradycardia model, the atrial fibrillation model, the left bundle branch block model, the right bundle branch block model, the premature atrial contraction model, and the premature ventricular contraction model, and the normal heart condition model to classify a true positive (TP) case in which cardiac abnormality is classified as the cardiac abnormality, a false negative (FN) case in which the cardiac abnormality is classified as normal, a false positive (FP) case in which the normal is classified as the cardiac abnormality, and a true negative (TN) case in which the normal is classified as the normal,
and calculates classification accuracy for at least one of the tachycardia model, the bradycardia model, the atrial fibrillation model, the left bundle branch block model, the right bundle branch block model, the premature atrial contraction model, and the premature ventricular contraction model based on a ratio of a combination of a numerical value of the true positive (TP) case and a numerical value of the true negative (TN) case to a combination of a numerical value of the true positive (TP) case, a numerical value of the false negative (FN) case, a numerical value of the false positive (FP) case, and a numerical value of the true negative (TN) case.

5 . The server according to claim 3 , wherein the artificial intelligence processor uses an open data set, the cardiac arrest model, and the normal heart condition model to classify a true positive (TP) case in which a cardiac arrest section is classified as the cardiac arrest section, a false negative (FN) case in which the cardiac arrest section is classified as a normal section, a false positive (FP) case in which the normal section is classified as the cardiac arrest section, and a true negative (TN) case in which the normal section is classified as the normal section, and calculates emergency state classification accuracy based on a ratio of a combination of a numerical value of the true positive (TP) case and a numerical value of the true negative (TN) case to a combination of a numerical value of the true positive (TP) case, a numerical value of the false negative (FN) case, a numerical value of the false positive (FP) case, and a numerical value of the true negative (TN) case.

6 . The server according to claim 3 , wherein the artificial intelligence processor classifies a true positive (TP) case in which an artifact signal is classified as the artifact signal, a false negative (FN) case in which the artifact signal is classified as a normal signal, a false positive (FP) case in which the normal signal is classified as the artifact signal, and a true negative (TN) case in which the normal signal is classified as the normal signal, and calculates artifact removal accuracy based on a ratio of a combination of a numerical value of the true positive (TP) case and a numerical value of the true negative (TN) case to a combination of a numerical value of the true positive (TP) case, a numerical value of the false negative (FN) case, a numerical value of the false positive (FP) case, and a numerical value of the true negative (TN) case.

7 . The server according to claim 6 , wherein, after the electrocardiogram signals are measured, the controller calculates the number of the measured electrocardiogram signals and compares the calculated number of the electrocardiogram signals with a threshold value to confirm a data reception state of the electrocardiogram signals.

8 . The server according to claim 6 , wherein the user terminal provides an analysis result related to the determined cardiovascular disease through a display.

9 . A sensor device for detecting a heart condition, comprising:
a biosignal monitor for measuring biosignals comprising electrocardiogram signals from a user and outputting monitoring information comprising the measured biosignals through an artificial intelligence encoder; and an artificial intelligence processor for extracting electrocardiogram signals comprised in the output monitoring information, extracting morphological information as feature information by converting the extracted electrocardiogram signals into a time-standardized image based on a pre-stored artificial intelligence machine learning algorithm, determining a plurality of cardiac abnormality type models using the extracted feature information, calculating classification accuracy for the determined cardiac abnormality type models, and determining a cardiovascular disease of the user using the determined cardiac abnormality type models and public cardiovascular disease data based on the calculated accuracy.

10 . The sensor device according to claim 9 , wherein the artificial intelligence processor simulates data traffic generated when measuring the electrocardiogram signals, and determines an operating state of the sensor device for detecting a heart condition based on the simulation.

11 . The sensor device according to claim 9 , wherein the artificial intelligence processor performs machine learning of the feature information to determine the cardiac abnormality type models as at least one model of a tachycardia model, a bradycardia model, an atrial fibrilla

Assignees

• Univ Korea Res & Bus Found

Inventors

• Kim Dong Joo
• Kim Young Tak
• Kim Hyun Ji

Classifications

• G16H40/63 Primary

  for local operation · CPC title

• G16H50/30

  for calculating health indices; for individual health risk assessment · CPC title

• G16H50/70

  for mining of medical data, e.g. analysing previous cases of other patients · CPC title

• G16H50/20 Primary

  for computer-aided diagnosis, e.g. based on medical expert systems · CPC title

• G16H40/67

  for remote operation · CPC title