Adaptive vector median filter for removal impulses from color images

This paper presents a new vector median filter that has the capability to detect impulses from color images. It skips the uncorrupted vector pixels in filtering color images, preserving image features undisturbed. The center vector pixel in the specified window is declared as corrupted if the detection mechanism finds the presence of the impulse in any one of the channels in the window. The detection mechanism incorporates the concepts of rank-conditioned median and center-weighed median filters in each image channel. To detect impulses, the rank-conditioned mechanism checks whether the central pixel is well within the ordered data set, while the center-weighted mechanism is used to select more appropriate local thresholds. The experimental results show the better performance of the proposed filter.