An ECG signal denoising method based on enhancement algorithms in EMD and Wavelet domains

This paper presents a new method based on enhancement algorithms in Empirical Mode Decomposition (EMD) and Discrete Wavelet Transform (DWT) domains for ECG signal denoising. Unlike the conventional EMD based ECG denoising methods that neglect a number of initial IMFs containing the QRS complex as well as noise, we propose a windowing method in EMD domain to filter out the noise from the initial IMFs without discarding them completely thus preserving the QRS complex. The comparatively cleaner ECG signal thus obtained from the EMD domain is employed to perform an adaptive soft thresholding in the DWT domain considering the advantageous properties of the DWT compared to EMD in preserving the energy and reconstructing the original ECG signal with a better time resolution. The performance of the proposed method is evaluated in terms of standard metrics by performing extensive simulations using the MIT-BIH arrhythmia database. The simulation results show that the proposed method is able to enhance the noisy ECG signals of different levels of SNR more accurately and consistently in comparison to some of the state-of-the-art methods.