Analysis of ECG Signal Classification Methods for Diagnosis of Heart Disease Using DWT and RBF Neural Network

Abstract

Heart diseases are common diseases which affects the human health. The arrhythmias, which might not be critically life-threatening but still, need attention and therapy to avoid deterioration. There is no fast and accurate method to analyse ECG signal. In this paper a new approach is designed for heartbeat classification based on DWT and RBF neural network. Discrete Wavelet transform will be applied separately to each heartbeat to extract static features. In addition, RR interval information like QRS points is computed to provide dynamic features. These two different types of features are concatenated as feature vector and a Radial Basis Function Neural Network classifier is utilized for the classification of heartbeats into one of different classes. The work is done on the data from two ECG leads and both decisions are fused for the final classification decision. It will be validated on the baseline MIT/BIH arrhythmia database for accuracy and different performance parameter. The clinical results are compared for diagnosis. The result of the work gives an automatic heartbeat classification.