Analysis of PCG Signal to Classify Various Heart Diseases

Abstract

Cardiac auscultation is a technique of listening to heart sounds. Any abnormality in the heart sound may indicate some problem in the heart. The abnormality in the heart sounds start appearing much earlier than the symptoms of the disease start showing. The phonocardiogram (PCG) signal is mainly recorded using an electronic stethoscope. In this study, the PCG signal i.e. the digital recording of the heart sounds has been studied and classified into three classes namely normal signal, systolic murmur signal and diastolic murmur signal. Various features have been extracted for the classification. The features extracted have been plotted against each feature to evaluate the features. It is seen that in some of the plots the three classes were completely separated. A total of 28 features have been extracted and then reduced to 7 features. The features have been selected on the basis of fisher discriminant ratio (FDR) feature reduction technique. The selected features are used to classify the signal into the pre-defined classes using various classifiers. The classifiers which have been used in this study were k-NN (k Nearest Neighbour), fuzzy k-NN and Artificial Neural Network (ANN). Highest accuracy of 99.6% is achieved using both k-NN and fuzzy k-NN as classifiers. Two new features have also been proposed for classification. Also, adaptive weighted FDR algorithm has also been developed to classify the signals. An accuracy of 96% is achieved using this algorithm.