Poornachandra S, Kumaravel N, Saravanan T, Somaskandan R
Member IEEE, Dept. of ECE, SSN College of Engineering, Anna University, Chennai, India.
Conf Proc IEEE Eng Med Biol Soc. 2005;2006:30-2. doi: 10.1109/IEMBS.2005.1616334.
WaveShrink plays pivotal role in analysis of complex bio signals. Wavelet analysis and WaveShrink are the established techniques of statistical research which have to be blended with the analysis of bio signals to provide a high degree of correlation between the desired signal and the observed signal. The multifaceted shrinkage function will obtain lurid responses to the signal perturbed to a very high degree by intransigent noise functions. It might be impossible to efface noise but its possible to ameliorate the recovery of the desired signal. Here we have obtained computationally efficient formula for computing the exact bias, variance, mean and risk. The former is the essence of any shrinkage function but latter is the essence of our paper. The purpose of our shrinkage function is to assuage the recovered signal to uphold our procedure for signal denoising and non-parametric regression. Our shrinkage function enjoys the same asymptotic convergence rate as the rest but offers obdurate resistance to pernicious factors disturbing the desired signal of interest.
