Wavelet transforms and denoising algorithms

Abstract

This paper addresses generalizations of the Donoho-Johnstone denoising method which is suited for additive Gaussian white noise and uses orthogonal wavelet transforms. We discuss the effects the choice of the wavelet transform has in wavelet shrinkage assuming the noise model to be Gaussian white noise. In addition we formulate some more general noise models that arise from real world data such as medical images and seismic data.