A novel algorithm for random-valued-impulse noise detection and removal using Chebyshev polynomial interpolation

This paper proposes an efficient way to remove random valued impulse noise and also the edge-preserving regularization of the henceforth obtained image. We follow a two phase mechanism where the noisy pixels in the interior regions are dealt with first and those in the edge regions later. Promising results were found even for Noise levels as high as 60% with the proposed algorithm.