Fast soft switching impulsive noise removal in color images

In this paper a novel soft switching filtering method intended for the reduction of impulsive noise in color images is presented. The noise detection algorithm is based on the concept of trimmed aggregated distances, which serve as a measure of pixel distortion. The output of the proposed filter is a weighted average of the central pixel and the mean of neighboring samples with the lowest degradation measure. The weighting coefficient is dependent on the degree of the central pixel corruption and in this way the filter output is able to adapt to the local contamination level. As a result, the undistorted samples are preserved and the pixels contaminated by impulsive noise are replaced by the mean of uncorrupted samples. The new filter outperforms existing methods and can be used in various applications in which the detail preserving noise reduction is required. An important feature of the new filtering design is its very low computational complexity which enables its application in real time denoising tasks.