Local self-adaptive fuzzy filter for impulsive noise removal in color images

This paper describes a new filter for impulsive noise reduction in color images based on a recently introduced family of vector filters with a good ability for preserving details. These filters use a reduced ordering of color vectors in order to detect and replace impulses. The proposed filter uses local statistics to adapt itself to simultaneously remove impulses and preserve image edges and details. The proposed filtering technique is robust and achieves a good balance between noise attenuation and detail reservation. In addition, it outperforms well-known vector filtering solutions and produces visually pleasing filtered color images.