Denoising of ECG signal based on improved adaptive filter with EMD and EEMD

Abstract

New improved methods for denoising Electrocardiogram (ECG) signal are proposed based on adaptive filter with Empirical mode Decomposition (EMD) and Ensemble Empirical mode Decomposition (EEMD). EMD and EEMD methods are used to decompose the ECG signal into intrinsic mode functions (IMF). Performance of traditional EMD based denoising methods improved by adaptively processing the IMF components which are related to ECG noise. Convergence issue in Least Mean Square (LMS) algorithm addressed by EEMD based adaptive algorithm. Block least mean square (ELMS) algorithm used with EMD and EEMD to improve the computational efficiency of adaptive processing. Proposed methods are applied on white Gaussian noise added ECG signal and real time ECG signals obtained from physionet MIT-BIH arrhythmia data base. Signal to Noise Ratio (SNR), correlation co-efficient and Mean Square Error (MSE) are used to measure and compare the performance of proposed methods with traditional EMD based methods. All the experiments done with MATLAB based coding. Results show that EEMD with ELMS algorithm performs better than traditional EMD based methods.