Wavelet based denoising algorithm of the ECG signal corrupted by WGN and Poisson noise

Electrocardiogram (ECG) signal has been widely used for heart diagnoses, but it always mixed with various noises. In this paper, we present a denoising algorithm that can be used for real-life ECG signal to increase signal to noise ratio for accurate diagnoses. Our new algorithm is based on a classical wavelet denoising process, which is taken from the Donoho et al. [5]. It is clearly appears that the P, T and R waves undergo distortions, which is due to the interference of the WGN. To overcome this drawback, our algorithm is to evaluate the corrupted WGN as well as Poisson noise and then to remove the noises interfering with R waves at the 3rd level detailed sequences and also to extract the QRS complex conjugate waveform of a ECG signal. In order to demonstrate our proposed method, our denoising algorithm is applying to a set of the MIT-BIH Arrhythmia Database ECG records that corrupted with a 10 dB white Gaussian noise. The universal threshold for the noise free reconstruction property is investigated and several simulations are carried out within MATLAB. The outcomes of the denoising performance of our algorithm are assessed by several measureable parameters, such as wavelet filters [Daubechies2 (DB2), Symlets (Sym2), Coiflet (coif2), DMeyer (demey), BiorSplines (bior2.2), ReverseBior (rbio2.2)]. The simulation results clearly show our denoising algorithm is efficiently and effective due to all the parameters are significantly improved in comparing with the other two different noise resources under the selected different mother wavelets.