Impulsive Noise Removal in Color Digital Images Based on the Concept of Digital Paths

In the paper a fast technique of impulsive noise suppression in color images is described. The proposed method is utilizing the concept of digital paths which link the central pixel of a filtering window with its boundary. Each path is assigned a cost and the minimum one is treated as a measure of the central pixel corruption. If the minimum cost of a link between the window center and its boundary is high, then the central pixel is treated as corrupted. The filter output is a weighted average of a pixel and its robust estimate, in which the minimum cost is used for the creation of a soft switching rule for the filter output. The performed experiments revealed excellent properties of the novel approach, especially in case of highly distorted images. The computational complexity of the proposed technique is low and it can be used in real time applications.