Artificial Neural Network based System using Transform Domain Methods for ECG Abnormality Detection

摘要

The Electrocardiogram (ECG) is a major tool for detecting abnormal condition of the heart. Coronary artery disease is reduction of blood supply to the heart muscles resulting in ischemic condition. The deposition of fat, cholesterol and calcium results in narrowing of the blood vessels. The narrowing of coronary arteries, if severe, results in heart attack. The various parameters obtained from ECG in time domain help to diagnose the coronary artery disease. In this paper effort is made to illustrate methods to detect ECG abnormality using Fourier Transform, Windowed Fourier Transform and the Wavelet Transform. Training and testing sets are prepared for Artificial Neural Networks using each of these transforms. Thus Artificial Neural Networks are used to detect the ECG abnormality. The suitability of each of these transforms for the ECG abnormality detection is also discussed.