Adaptive median based multi-shell-order-statistics filters for the removal of impulsive noise

The multistage and multi-shell median filters have been shown to be effective in preserving fine image structures as well as in the suppression of noise. It has been shown that the multi-shell median filter can be modified to describe a set of median based multi-shell order-statistics filters, each with a different noise removal characteristic. In the paper, an adaptive median based multi-shell order-statistics filter is presented. The adaptive algorithm varies the statistical properties of the filter by updating the rank of the complementary order-statistics operators of each shell according to the input noise density. Its performance is compared to the standard 2D median and multistage median filters with favourable results.<
>