An efficient adaptive algorithm for removal of impulse noises

An efficient algorithm Adaptive Min-Max Average Filters (AMMAF) for the removal of impulse noises is proposed in this paper. AMMAF divides all pixels of the corrupted image into three types and has two phases denoising with adaptive small filters, which preserve most original signals and restore more details. Experiments results show that compared with state-of-the-art salt & pepper Denoising algorithm, AMMAF performs better and is also less time-consuming, especially for the high-density noise images.