Comparative Analysis of Shannon and Hilbert Envelograms for Heart Sound Segmentation

Abstract

The objective of this study is to compare the performances of envelope extraction algorithms used to segment heart sounds under normal conditions as well as during valvular disorders. Two different approaches were followed to implement the extraction of envelope of the first and second heart sounds-Normalized Average Shannon Energy (NASE) and Hilbert Transform method. Envelope extraction using the above stated algorithms were performed on raw as well as signals denoised using Discrete Wavelet transform (DWT) and the time duration parameters associated with each heart sound were computed after segmentation. Based on the results obtained, the efficiency of each algorithm was studied and their advantages and disadvantages were compared.