Method For The Detecting Electrocardiogram Anomalies And Corresponding System

Abstract

A heart-rate associated with a heartbeat signal is determined. A transform is selected based on the determined heart-rate associated with the heartbeat signal and a reference heart-rate associated with a dictionary of a sparse approximation model. The transform is selected independent of other factors associated with generation of the heartbeat signal. The selected transform is applied to the dictionary of the sparse approximation model, generating an adjusted dictionary of the sparse approximation model. Anomalous heartbeats in the heartbeat signal are detected using the adjusted dictionary of the sparse approximation model.

Description

technical field [0001] The present disclosure relates to techniques for detecting anomalies in electrocardiogram (ECG) signals, such as those acquired by wearable devices used for health monitoring. Background technique [0002] Anomaly detection in ECG signals refers to the problem of finding patterns that do not conform to the expected behavior of the monitored signal. These abnormal patterns can be translated into salient information, such as abnormal heart activity (arrhythmia), malfunctions in the sensing process, etc. [0003] Wearable devices can continuously acquire, store, and periodically transmit physiological signals that are typically processed by a host device in real time. The morphological characteristics of ECG signals may show significant variation for different users, so general methods may not perform well on wearable devices. For example, the ECG morphology may be modified each time the wearable device is repositioned or subjected to unexpected movemen...

AI Technical Summary

Problems solved by technology

Morphological characteristics of ECG signals may show significant variation for different users, so general methods may not perform well on wearable devices

Images