Mitigation of noise and interference in ECG signals with Savitzky-Golay least squares polynomials and Discrete Wavelet Transform

ECG signal is mostly used in biomedical research and for clinical diagnosis. Recorded ECG signal often contains a noise and interference with non-stationery properties. Electrode noise, electronic noise and motion artifacts are most common type of random noises which are recorded in the ECG signal. One of the major problem in biomedical signal acquisition is to separate small input signal from noise and disturbance caused by 50 Hz power supply. The main objective of this paper is the performance analysis of two adaptive filters namely Savitzky-Golay (SG) filter with and without smoothening and Discrete Wavelet Transform (DWT) for noisy ECG stimulus. Results are encouraging. The Mean Square Error (MSE) reduced to 5.07, 3.72, and 1.101 with Discrete Wavelet Transform (DWT), Savitzky-Golay (SG) Filter and Savitzky-Golay (SG) Smoothing Filter respectively. Hence SG Filter with smoothing is very useful for denoising ECG signals in medical diagnosis applications.