Heart auscultation is a fundamental tool in the diagnosis of heart diseases. Although nowadays it has been less focused due to the advent of ECG and echocardiography; still, there are some cardiac defects that are best detected by heart sounds. The mechanisms that generate cardiac sounds are complex and there is, so far, no general consensus as to the contribution made by various mechanical cardiac events to the formation of the single components. However, it is agreed by most investigators that the mechanical activity of the heart including the blood flow, vibrations of the chamber walls and opening and closing of the valves are the major reasons for generation of the PCG signal.

The major audible components of a normal cardiac cycle are short beats which are recognized as the primary components S1, S2.

 

 

S1, which is usually the longest and loudest component of the PCG signal, occurs at the onset of the ventricular contraction. It lasts for an average period of 100ms−200ms and its frequency components lie in the range of 25Hz−45Hz. The second heart sound component, S2, is heard at the end of the ventricular systole and lasts about 0.12s with a frequency less than 150Hz.

PCG signal is accompanied by noise components introduced by various sources such as ambient noise, respiratory and/or thoracic muscular noise, non-optimal recording sites, and weak sounds; however, providing an accurate subject recognition demands considerable reduction of the noise contribution. Investigating some of the PCG’s characteristics such as its ability to discriminate individuals as well as its stability over a period of time, demonstrated the possibility of utilizing PCG signal for human recognition. Our group is currently investigating feature extraction using wavelet analysis to employ PCG signals in biometric recognition. We are also looking into fusion mechanisms of ECG and PCG into a multi-modal biometric system.