Three-Phase Distance Matrices Filter
Wei-Chun Hsu, Yuan‐Chen Liu, Kuan-Yu Pan · 2009
In this paper, a new noise filtering, three-phase distance matrices filter (TPDMF), for restoration of corrupted image by impulsive noise is presented. Whether images were attacked by random valued impulse or constant valued impulse, this method can restore corrupted images fast and effectively without iterative procedure. For avoiding oversmoothing after filtering, TPDMF uses distance matrices to preserve edge and detail of image. TPDMF consists of threephase: impulse detection, impulse verification and impulse noise cancellation. Impulse detection phase will pick up alleged pixels. It will be verified that it is impulse or not in second phase. Finally, the impulse noise will be replaced by estimated value. This algorithm uses the concept of distance matrices that be consist of median distance matrices and POI distance matrices produced by each pixel. Results from computer simulation show that TPDMF provides a signification improvement over many other existing techniques.