Expert system for phonocardiographic monitoring of heart failure patients based onwavelet analysis

An expert system was developed basically for the phonocardiographic (PCG) monitoring in the coronary care unit and at home. The aim of the study was to examine the performance of wavelet transformation in a heart failure group with noisy environment. For detection and interpretation of heart sounds and murmurs two analysis methods were done. As first, with the waveform similarity overlap-and-add (WSOLA) algorithm, a better separation of sounds was achieved for determination of the gold standard to the cardiologist interpretors. For the time-frequency analysis of PCG signals in a special clinical settings (PCG monitoring of 52 heart failure patients (NYHA III-IV) with ejection fraction<0.35), various wavelets were compared to an autoregressive (AR) model to determine the better model. As a result the Morlet wavelet wave showed the best performance, so it will be used for the future in our PCG data-, and knowledge-base