Research on ECG Signal Denoising by Combination of EEMD and NLM

In modern medicine, it is necessary to collect signals to determine human's health, such ECG signal, a fluorescent signal obtained by Polymerase Chain Reaction (PCR) in medicine, and the like. In this article, in allusion to the myoelectric interference, baseline drift and common interference caused by machine power and external environment, A new signal denoising method based on the combination of Ensemble Empirical Mode Decomposition and Non-Local Mean is put forward. The method decomposed the ECG signal by EEMD to obtain a set of IMFS, then calculate the sample entropy of each component, and distinguish the signal dominant component between noise dominant component, and then denoise the noise dominant component. by Non-Local Mean to obtain a new component, and finally reconstructs the new component and the signal dominant component to gain a denoised signal. In experimental verification, the new algorithm is compared with wavelet threshold denoising, EEMD-wavelet threshold denoising and normal Non-Local Mean denoising. According to the experimental results we draw the conclusion that the new method can remove the noise influence effectively on the Signa Noise Ratio and Mean Square Error. Both are better than several other algorithms.