A regional multimodulus algorithm for blind image deconvolution

We propose an efficient scheme for blind image restoration that takes advantage of recent advances in blind algorithms used for equalization of communication channels. It is based on a blind adaptive algorithm that interprets images with more than two grayscale levels as images with only two grayscale levels. In order to take a greater advantage of the adaptive scheme, we also propose some changes in the implementation of the algorithm. The performance of our proposal is illustrated by means of simulations, where we show its advantages with respect to an existing image deconvolution scheme based on the two-dimension constant modulus algorithm.