A Study of Chosen an Optimum Type of Wavelet Filter for De-Noising an ECG signal

Abstract

Among various biological signals for the diagnosing of cardiac arrhythmia, Electrocardiographic (ECG) signal is the most significant one. The interesting challenge is an accurate analysis of the noisy ECG signal. Prior to accurate analysis, these signals need for de-noising to remove these unwanted noises in the signal to get an accurate diagnosis. In order to get the best de-noising results, it should have an accurate decision about the filters that we deal with for de-noising the signals. So, in this paper we present a study for choosing the optimum wavelet filter for de-noising the electrocardiograph (ECG) signal. Signals were stored as a one-dimensional matrix and series of procedure were performed to reduce the noise. The wavelets filters were chosen that very close to the original signal after applied a random-noises to the ECG signals to get familiar with the possible noise that can the signal affected with it. Also, estimation the most standard wavelet families namely Symlets, Coiflet, and Daubechies with different methods of threshold and decomposition levels were done. The purposes of this study to conclude the convenient wavelet functions in decomposition, the de-noising and the reconstruction, the method of the threshold, and the optimal decomposition level of the wavelet.