Signal processing techniques for phonocardiogram de-noising and analysis

Abstract

Explores phonocardiogram de-noising techniques, including wavelet de-noising. Optimised wavelet de-noising, wavelet packet de-noising, matching pursuit technique, and averaging. Optimised wavelet de-noising performed slightly better than other methods, and is recommended to be used in conjunction with averaging. Also explores different methods of extracting features from the de-noised phonocardiogram and classifying it. These include phase space diagrams, HT diagrams, instantaneous signal parameter extraction and phase synchronisation between the ECG and PCG; investigations were limited by the quantity and quality of data available