A Novel Empirical Mode Decomposition Denoising Scheme

This paper presents a novel and fast scheme for signal denoising by using Empirical mode decomposition (EMD). The EMD involves the adaptive decomposition of signal into a series of oscillating components, Intrinsic mode functions(IMFs), by means of a decomposition process called sifting algorithm. The basic principle of the method is to reconstruct the signal with IMFs previously selected and thresholded. The denoising method is applied to four simulated signals with different noise levels and the results compared to Wavelets, EMD-Hard and EMD-Soft methods.