Image de-noising via overlapping wavelet atoms

This paper focuses on a novel approach for image denoising: WISDOW (Wavelet based Image and Signal De-noising via Overlapping Waves). It is based on approximating any singularity by means of a basic one in a wavelet domain. This approach allows us to reach some interesting mathematical properties along with good performances in terms of both subjective and objective quality. In fact, achieved results are comparable to the best wavelet approaches requiring a low computational effort and resulting completely automatic.