An investigation on the performance analysis of ECG signal denoising using digital filters and wavelet family

Abstract

Electrocardiography is a technology used to identify the abnormalities in heart and noise free ECG data is often required for correct medication of cardiac disorders. Generally ECG signals are contaminated by noise and human artifacts during data acquisition. The denoising signal plays a major role in medical field. Electrocardiogram (ECG) signals represent important characteristics for diagnosing the disease or how the treatment works on the heart, which makes it necessary to design filters to weaken and eliminate these noises. This paper describes the denoising of ECG signal from baseline wander noise using digital filters and wavelet transform. The function of the filters has been tested on different cardiac signals. The results show that wavelet transform has the best performance in denoising ECG signals than digital filters such as IIR (Infinite Impulse Response) notch and window based FIR (Finite Impulse Response) filters. Finally the performance of the wavelet based approach is evaluated with SNR (Signal to Noise Ratio) value, PSNR (Peak Signal to Noise Ratio) value, Mean Square Error (MSE) value and Correlation Coefficient (CC) value and compared among various wavelet families. All simulations are carried out using MATLAB.