Power line noise and baseline wander removal from ECG signals using empirical mode decomposition and lifting wavelet transform technique

The potency of the intuitive empirical mode decomposition in conjugation with the efficient and fast lifting wavelet transform in discarding powerline noise and baseline wander is studied in this paper. The Empirical Mode Decomposition (EMD) disintegrates the noisy ECG signal into a band of intrinsic mode functions with the noise remaining confined within a few modes. The sum of the noisy modes corresponding to the Power Line Interference (PLI) and the baseline wander are separately fed to a multi-level lifting wavelet transform where the coefficients corresponding to the noise are set to zero using a hard threshold scheme. A periodic extension method is used in each case to take care of the border effects. The noise free modes are reconstructed using the inverse Lifting Wavelet Transform (LWT). The processed intrinsic modes are added to the remaining unprocessed modes to reconstruct the final noise free signal. The QRS complex preserved using an R-peak location mechanism is finally added to the signal. The prowess of the method has been tested by calculating the increase in the signal-to-noise ratio as well as the cross-correlation coefficient values between the clean input message signal and the denoised output signal at different input SNR values for various ECG records available online at MIT-BIH Arrhythmia database. The results have been compared various other methods to establish the superiority of the proposed method.