An adaptive denoising method based on local mode estimation

An adaptive method is presented in this paper to cope with Gaussian noise through local structure estimation. Even regions and uneven regions are distinguished firstly, then each uneven region is further segmented into two subregions by fuzzy c-means clustering algorithm and Fisher discriminant analysis. Finally, linear approximation is used to estimate the intensities of each region. With the introduction of local structure analysis, Gaussian noise is reduced effectively with edges preserved very well. Experimental results also demonstrate its effectiveness in terms of medical images enhancement, while it is also applicable to other similar images that have many tiny details