Heart Sounds Parameter Extraction for Automatic Diagnosis

Abstract

For efficient heart inspection with an automatic and noninvasive system it is imperative to collect different signals from human body. In the sequence of the development of a standalone Portable Decision Support System for Heart Failure Detection and Medical Diagnosis it was decided to record heart sounds to help in diagnosis process. However, automatic identification and parameter extraction of heart sounds deals with several challenges, such as, noise sources and low level signal pressures at microphone, as well as, abnormal sounds that overlaps normal signals. This paper discusses phonocardiography and its signals, as well as, signal processing techniques able to extract important parameters and indicators from heart sounds.