Quantitative study on the selection of wavelet functions for the de-noising of ECG signal

Abstract:

To quantitatively evaluate the de-noising effects of different wavelet functions on the electrocardiogram(ECG) signal，a noisy ECG model was constructed as an experimental standard signal，which was processed by orthogonal wavelet transform and different thresholding methods to remove high-frequency noise. The de-noising effects were measured by Signal to Noise Ratio(SNR) and the shape of waveforms. It was demonstrated that when the SNR after denoising approached to that of the standard signal，the de-noising effect was optimal. Thus，the de-noising schemes with the less signal distortion and the higher SNR were obtained. Meanwhile，the wavelet function appropriate for the decomposition and reconstruction of ECG signal was determined. Verified by MIT-BIH database，the results of this study can effectively remove high-frequency noise in ECG signal.