High density impulse denoising by a novel adaptive fuzzy filter

Traditional median filters perform well in restoring the images corrupted by low density impulse noise, but fail to restore highly corrupted images. Conversely, the advanced adaptive median filters are capable of denoising high density impulse noise but the image details are compromised significantly. In this paper, a new adaptive fuzzy median filter is presented to provide optimum detail preservation along with very high density noise removal. The novelty of this research work comes from two directions. Firstly, we used a triangular fuzzy membership function to determine the level of corruption at each pixel that consequently ensures the replacement of noisy pixels according to the extent of corruption. Secondly, we exploited fully adaptive and automatically adjustable threshold value to provide ease of computation. Experimental results show that the proposed filter outperforms other conventional and advanced filters in terms of both denoising and fine detail preservation of highly corrupted images.