Analysis of de-noising techniques of non-stationary ECG signal based on wavelet and PSO optimized parameters for Savitzky golay filter

Abstract

The bioelectric signals originate from different organs of human body. Out of these ECG is one of the important bioelectric signals for our concern and for investigation of various heart ailments. ECG demonstrates graphical rendition of activity of the heart and by analyzing the ECG signal several heart diseases can be identified. For analysis of ECG signal, it must be noise free. There are several techniques for filtering out the noise from ECG signal. It is imperative to find out a best possible method for ECG Signal de-noising, which may be investigated by finding and comparing the signal-to-signal-plus-noise ratios (SSNR) and root-mean-square deviations (RMSD). By considering a mother wavelet with different levels and threshold, the values of SSNR and RMSD are calculated and compared to achieve a best possible result. Similarly, by utilizing different frame lengths and polynomial orders, the S-G filter is employed to find SSNR and RMSD. The results of these methods are also compared. Finally, the orders and frame lengths of S-G filters can be obtained by optimizing with the help of particle swarm optimization (PSO) technique.