A new framework for wavelet based analysis of acoustical cardiac signals

Cardiac auscultation is one of the classical methods used by the physicians for diagnosis of cardiac abnormalities. There have been attempts to develop complex algorithms for automated diagnosis based on heart sounds. It is however, observed that developmental work towards integrated automated auscultation system targeted towards mass screening has been relatively low. In this paper, we propose a new framework for analysis of acoustical cardiac signals which is suitable for development of point of care cardiac healthcare products. Wavelet analysis has been employed for denoising of the signal which enables the systems applicability in a rural health centre. The new framework also effectively focuses on selection of best wavelet basis through entropy measures and reducing the time-frequency decomposition cycles.