Heart sound analyzer

Abstract

Despite advancement in medical technology throughout the years, cardiovascular diseases remain the leading cause in worldwide deaths. It is extremely vital to detect heart illnesses before the onset of an attack. There are many types of scans, ranging from MRI (magnetic resonance imaging) to coronary angiography. However, this scans are expensive and very often invasive tests.

Heart sounds are non-stationary signals which are caused by the sudden closing of heart valves at different stages of cardiac cycle. They carry information on the physiological condition of the heart. As such, heart auscultation, which is defined as listening to the heart sounds, is another invaluable tool in assisting medical practitioners and physicians in diagnosing the patients.

This project makes use of signal processing techniques, segmentation and filtering to perform the automatic detection of heart sounds. The algorithm is written on the software MATLAB R2009b and primarily uses wavelet transform to detect the location of fundamental heart sounds and heart murmurs.