The adaptive bilateral filter (ABF) for improved sharpening and elimination of noise. ABF sharpens an image by increasing the slope of the edges without producing overshoot or undershoot. It is an approach to enhancing sharpness that is fundamentally different from the unsharp mask (USM). This new approach to slope restoration also differs significantly from previous slope restoration algorithms in that the ABF does not involve edge detection or its orientation, or the extraction of edge profiles. In ABF, the slope of the edge is improved by transforming the histogram through a range filter with adaptive displacement and width. The ABF is able to smooth out noise while improving the edges and textures in the image. The parameters of the ABF are optimized with a training procedure. The ABF restored images are significantly sharper than those restored by the bilateral filter. Compared to a USM-based sharpening method, the optimum mask (OUM), the restored ABF edges are as sharp as those shown by the OUM, but without the halo artifacts that appear in the restored OUM image. In terms of noise elimination, ABF also outperforms the bilateral filter and OUM. We show that ABF works well for both natural and text images.

In terms of noise elimination, the conventional linear filter works well in soft regions, but significantly erases the edge structures of an image. Much research has been done on noise elimination that preserves the edges. One of the main efforts in this area has been to use the classification order information. Due to the lack of spatial ordering, sorting order filters generally do not suppress gaussian noise optimally. In more recent years, Tomachi et al. Y Smith et al. They proposed a new approach to eliminate noise. Although their algorithms were developed independently, and respectively called "SUSAN" filter and "bilateral filter", the essential idea is the same: to reinforce the geometric proximity in the spatial domain and the similarity of gray values in the range in the de-noising operation.