EMD interval thresholding denoising based on correlation coefficient to select relevant modes

Empirical mode decomposition(EMD) has been proved that it is an extraordinary useful nonparametric signal denoising, which adaptively breaks down any signal into oscillatory components called intrinsic mode functions(IMFs). For EMD-based denoising, how to select relevant extracted modes and how to select threshold for denoising are two main questions to resolve urgently. This paper introduces a new EMD-based denoising, tremed EMD-corIT, which combines an estimation of correlation coefficient and the EMD interval thresholding denoising (EMD-IT). The selection of relevant modes is dependent on the correlation coefficient between the IMFs and the original signal. Numerical simulation test were carried out to evaluate the effectiveness of this method. Other traditional EMD denoising, such as correlation-based EMD partial reconstruction, EMD direct thresholding (EMD-DT) are also investigated in this paper to provide a comparison. Simulation results show its superior performance over other traditional EMD denoisings in whole.