Improved Image Denoising Filter using Low Rank & Total Variation
Better diagnosis of disease is possible only with the better microscopic images. To do so images of the affected area are captured and then noise is removed to obtain accurate diagnosis. Many algorithms have been proposed till date. But they are capable of removing noise only in spatial domains so this paper tries to overcome that by combining low rank filter and regularization. If we only reduce noise in spatial or spectral domain, artefacts or distortions will be introduced in other domains. At the same time, this kind of methods will destroy the correlation in spatial or spectral domain. Spatial and spectral information should be considered jointly to remove the noise efficiently. Low rank algorithms are good as they encloses semantic information as well as poses strong identification capability.
