Modified switching median filter for impulse noise removal

In this paper, a modified switching median filter is presented for noise reduction in color images that are corrupted with impulse (salt and pepper) noise. It is a two-phase noise detector: in the first phase, we use the adaptive vector median filter detection method to identify pixels that are likely to have been corrupted by noise (as noise candidates); in the second phase, these noise candidates are judged by using four one-dimensional Laplacian operators, which allow edge pixels to be preserved. Extensive experiments show that the proposed method outperforms many existing vector filters in terms of filtering performance. In particular, the proposed approach can effectively preserve thin lines, fine details, and image edges.