ECG denoising using adaptive selection of IMFs through EMD and EEMD

In this paper the removal of artifacts from ECG signal has been done using two different algorithms, Empirical Mode Decomposition (EMD) and Ensemble Empirical Mode Decomposition (EEMD). The mode mixing problem of EMD is alleviated by adding white noise and ensembling the intrinsic mode functions in case of EEMD. Different types noises were added to the ground truth (GT) signal and comparison has been done in terms of signal to noise ratio(SNR db). MIT-BIH database was used to acquire the ECG signal and noise signals. Correlation coefficients were calculated among different combinations of IMFs and ground truth signal. Reconstruction of denoised ECG signal has been done on the basis of the correlation coefficients. The result shows that the performance of EEMD algorithm is superior than the EMD algorithm.