Comparison of time-frequency analysis performance on hearts and murmurs

Abstract

Time-frequency analysis is one of the methods to deal with non-stationary and time-varying signal such as heart sounds and murmurs. Wigner-Ville distribution (WVD) is a popular tool to obtain time-frequency representation of signal. Windowed Wigner-Ville distribution (WWVD) and smooth windowed Wigner-Ville distribution (SWWVD) are the improvement method to overcome the problem faced by WVD. Determination of parameter setting of WWVD and SWWVD eliminates the cross-terms and improve time-frequency representation. The accuracy of time-frequency representation of simulated heart sounds and murmurs are compared and determined based on the mainlobe width, peak-to-sidelobe average ratio and signal-to-interference ratio. It is found that the most accurate time-frequency representation can be obtained using the SWWVD.