A feature extraction algorithm based on wavelet packet decomposition for heart sound signals

Abstract

In this paper, a feature extraction algorithm based on the wavelet packet decomposition (WPD) method was developed for the heart sound signals. Feature vectors obtained were used to classify the heart sound signals into physiological and pathological murmurs. The classification using a neural network method indicated a 85 percent accuracy. This could be an effective assistance for medical doctors to make their final diagnoses.