This study presents a fuzzy-based approach using wavelet transforms to effectively denoise ECG signals, which is essential for accurate feature extraction in cardiac diagnostics and other applications.
Electrocardiogram (ECG) is a vital biomedical signal for diagnosing heart diseases, but now it has many other applications like stress recognition, biometric recognition etc. but ECG signal gets noisy from various sources like as muscle noise, electrode artifacts, baseline drift noise and respiration. As wavelet transform shows a good performance in de-noising the ECG signal, however the selection of appropriate mother wavelet functions and number of wavelet decomposition levels is still an issue to remove the various kinds of noises from the input signal. It is essential to denoise the ECG signal to get appropriate features of ECG signal.This research work analyze and compare the removal of noise and distortion in ECG signal using five wavelets(Daubechies, Coiflet, Haar, Biorthogonal and Symmlet) with four thresholding rules(SURE, Hybrid,Universal and Minimax) and various decomposition levels using Fuzzy Inference System .The parameters used for performance analysis are Signal to Noise Ratio,Mean Square Error(MSE) and variance.
Goel et al. (Sat,) studied this question.