Quantification of non-deterministic events in pathological PCG signals using Continuous and Packet Wavelet Transforms

The Phonocardiogram signal is a combination of contraction and relaxation of atria and ventricles. Valve movements play an important role in the medical field by providing the clinician with diagnostic and prognostic information. The energy associated with the signal can be useful in quantification of various non-deterministic events such as cavitation in mechanical heart valve patients, mitral regurgitation, pulmonary stenosis and murmurs. As the PCG signals are complex and non-stationary, the energy calculated in time and frequency domain is difficult to analyze. This paper analyzes the importance of using wavelets in non-deterministic energy calculation using two versions of wavelet transforms: the Continuous Wavelet Transform (CWT) and Packet Wavelet Transform (PWT). The analysis is obtained by using large number of orthogonal and bi-orthogonal wavelets in finding the difference between the deterministic and total energy, which leads to the non-deterministic energy of the signal. The performance of CWT and PWT in non-deterministic energy calculation is evaluated. The results obtained are analyzed and discussed.