Enhancement of ECG signal using EMD and hurst-based mode selection technique

An ECG signal contains important information required for diagnosis and analysis of heart diseases. So if there is noise induced in an ECG signal then scrutinizing of that signal for pathological, anatomical and physiological aspects goes worthless. Noises can be introduced by various sources, but a common source for high frequency noise is due to forces acting on the electrodes. In this paper noise removal from ECG signal is based on empirical mode decomposition (EMD) and a set of intrinsic mode functions (IMF) is obtained. The main contribution here is adopting Hurst exponent in the selection of IMFs to reconstruct the cardiac signal. This EMD and Hurst-based (EMDH) approach is evaluated in cardiac signal enhancement experiments considering environmental noises with different indices of non-stationarity. Simulation here is done on the MIT-BIH database to evaluate proposed algorithm. Experiments show that the presented method offers good results to detect characteristic waves and remove noise from the ECG signal.