Wavelet based bank of correlators approach for phonocardiogram signal classification

Abstract

This paper presents a methodology for detecting all the components of the phonocardiogram (PCG) signal based on a time-scale map obtained from a proposed wavelet based bank of correlators, without the aid of any additional reference signal to provide cardiac phase information. The cardiac phase information is obtained during the classification phase. The novel approach, namely the wavelet based bank of correlators approach uses the Morlet wavelets as the correlating filters. In this paper, the rationale for the choice of Morlet wavelets as the correlating filter is also developed. A complete scheme for detection and classification of PCG signals based on a simple perceptron is also presented.