Heart sound segmentation based on homomorphic filtering

Abstract

Phonocardiography, the digital recording of heart sounds, is becoming increasingly popular as a primary detection system for diagnosing heart disorders and is relatively inexpensive. The electrocardiogram (ECG) is often used when recording the phonocardiogram in order to help identify the systolic and diastolic components. In this study, a heart sound segmentation algorithm has been developed which automatically separates the heart sound signal into these component parts. 100 patients with normal and abnormal heart sounds were studied. The algorithm uses homomorphic filtering to produce time-domain intensity envelopes of the heart sounds and separates the sounds into four overlapping parts: the first heart sound, the systolic period, the second heart sound and the diastolic period. The performance of the algorithm was evaluated using 14,000 cardiac periods from 100 digital phonocardiographic recordings, including normal and abnormal heart sounds. In tests, the algorithm was over 93 percent correct in detecting the first and second heart sounds. The automatic segmentation algorithm presented uses wavelet decomposition and reconstruction to select a suitable frequency band for envelope calculations and has been found to be effective in segmenting phonocardiogram signals into four component parts without using an ECG reference.