Wavelet Transform Based Estimation of 1 Dimensional Signal

Abstract

The key idea of this manuscript is denoising of noisy biological signals. For this wavelet thresholding technique is suggested. To eliminate the noise existing in the signal, mixed thresholding function is considered which is the median of Hard, Soft and Garrote functions. The mixed thresholding function is applied by degraded white gaussian noise Electrocardiogram signal. Two methods that are used to calculate the threshold value is FDR technique and Visu shrink technique. The outcomes of mixed functions are compared with remaining functions using Signal to Noise Ratio (SNR) and Mean Square Error (MSE). It is obvious that the mixed function performs superior than remaining functions using Visu shrink technique and performs better than only Hard function using FDR technique.