Optimal selection of wavelet thresholding algorithm for ECG signal denoising

In this paper we examine denoising performance of four wavelet thresholding algorithms i.e., Universal, Rigrous SURE, Minimax with hard and soft threshold, and Neighbourhood based threshold on synthetic and real ECG signal. We apply the Stationary Wavelet Transform to decompose ECG signal into wavelet domain. Performance analysis was performed by evaluating Mean Square Error (MSE) and visual inspection over the denoised signal from each algorithm. The experimental result shows that Universal hard thresholding gives the best denoising performance on synthetic ECG signal. This result is consistent with the experiment on real ECG signal. The result shows that soft threshold not always gives better denoising performance; it depends on which wavelet thresholding algorithm was choosen. The use of Neighbourhood based thresholding on synthetic and real ECG signal shows over smooth denoised signal.