Segmentation of Pathological Heart Sound Signal Using Empirical Mode Decomposition
Abstract
The Phonocardiogram (PCG) is the graphicalrecording of acoustic energy produced by the mechanicalactivity of various cardiac. Due to the complicated mechanismsinvolved in the generation of in the PCG signal, it is consideredas multicomponent non stationary signal. Empirical modedecomposition (EMD) allows decomposing an observedmulticomponent signal into a set of monocomponent signals,called Intrinsic Mode Functions (IMFs). The goal of this paperis to segment some pathological HS signals into the murmursrelated to cardiac diseases. EMD approach allows toautomatically selecting the most appropriate IMFscharacterizing the murmur using the noise only model. Real-lifesignals are used in the various cases such as Early AorticStenosis (EAS), Late Aortic Stenosis (LAS), MitralRegurgitation (MR) and Aortic Regurgitation (AR) to validate,and demonstrate the effectiveness of the proposed method.