Anomaly Detection in Electrocardiogram Signals using Autoencoders

Publications

Anomaly Detection in Electrocardiogram Signals using Autoencoders

Author : Dr Pavithra M R

Year : 2024

Publisher : Institute of Electrical and Electronics Engineers Inc.

Source Title : 2024 Asian Conference on Intelligent Technologies, ACOIT 2024

Document Type :

Abstract

One potential use of autoencoders in healthcare is the identification of anomalies in electrocardiogram (ECG) readings. The primary objective of this research is to design a reliable system that can detect irregularities in electrocardiogram data automatically. The goal is to improve the efficiency and accuracy of anomaly detection using autoencoders so that cardiovascular problems may be diagnosed and treated early. The objective is to prove that the suggested approach can correctly detect a variety of abnormalities, such as arrhythmias and other cardiac abnormalities, using electrocardiogram (ECG) data by conducting extensive experiments and validations. This study aims to provide a dependable tool for healthcare practitioners to evaluate ECG data effectively and quickly detect possible health issues, therefore contributing to the improvement of medical diagnostics. The goal of our effort is to make clinical decision-making in cardiac care settings more effective and to improve patient outcomes. The PTB Diagnostic ECG Database yielded five results, ranging from 0.57 to 0.92, for each of the five patients in the sample.