A technique for image restoration based on recursive processing and error correction

A technique for restoration of images corrupted by mixed noise is presented. The proposed approach aims at removing the different noise components by means of a recursive process that minimizes filtering errors. The method combines simplicity and effectiveness. No complicated tuning of system parameters is necessary. A very accurate restoration of data corrupted by mixed Gaussian and impulse noise can be obtained. Experimental results show that the proposed method can largely outperform other techniques in the literature.