Heart sound localization in chest sound using convex-hull algorithm

This paper presents a technique to estimate heart sound localization in chest sound. In the proposed technique, chest sound is divided into frames and for each frame various types of acoustic features are extracted. The acoustic features used in this work are energy, log-energy and entropy with their smoothed versions. The heartbeat locations are detected by an algorithm known as convex-hull algorithm. It uses the features extracted from chest sound and pre-defined threshold in heartbeat detection process. The experimental results show that the proposed algorithm achieves 98-100% correct rate for detection of heartbeat locations.