Three levels discrete wavelet transform elliptic estimation for ECG denoising
Amel Baha Houda Adamou‐Mitiche, Lahcène Mitiche, Hilal Naimi · 2016
Electrocardiogram (ECG) signal plays an important role in the primary diagnosis, prognosis and survival analysis of heart diseases. Several noise types are sources of ECG signal corruption such as electrode movement, strong electromagnetic effect and muscle noise. The principle of techniques based wavelet depends on shrinking the wavelet coefficients in the wavelet domain. These techniques prove their efficiency for their ability to capture the energy of a signal in few energy transform values when removing the noise. In this paper, a new denoising approach based on the combination of three levels discrete wavelet transform and Elliptic filter estimation is presented. To validate and conclude to the efficiency of our approach, a physical signal is processed as well by our method as with two recent denoising techniques, namely the discrete wavelets transform thresholding (hard or soft). Based on two important measures, the quality of our method is confirmed.