Denoising Via Block Wiener Filtering in Wavelet Domain
Vasily Strela · Birkhäuser Basel eBooks · 2001
In this paper we describe a new method for image denoising. We analyse statistical properties of the wavelet coefficients of natural images. It turns out that there is a strong local covariance structure introduced by the edges. We suggest a model for this covariance which allows us to estimate it from the noisy image. Then Wiener filter is employed in order to remove the noise. We compare our approach to other noise removal techniques. Wienerwavelet denoising produces superior results both visually and in terms of mean square error. These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.