An adaptive content based closer proximity pixel replacement algorithm for high density salt and pepper noise removal in images

Abstract

An Adaptive Content based Closer Proximity Pixel Replacement algorithm for the removal of high density salt and pepper noise in images is proposed. The algorithm uses decision tree to identify and correct the pixels of the image is noisy or not. The algorithm finds Euclidean distance between the processed pixel and the number of non-noisy pixels inside the current processing kernel. The algorithm requires only two non-noisy pixels to be present in kernel for the algorithm to operate. The faulty pixels are replaced only by the median of pixels that occurs more frequently in the current processing kernel based on the Euclidean distance. The algorithm increases the window size by two when there are no non-noisy pixels in the current processing kernel. The proposed algorithm was compared with 16 standard and existing algorithms derived from recent literatures. Exhaustive experiments on standard database images suggest that the algorithm exhibit excellent noise suppression and good information preservation characteristics even at very high noise densities.