Parametric Image Restoration Using Consensus: An Application to Nonstationary Noise Filtering

Image quality gets affected by unavoidable degradations. Several techniques have been proposed based on a priori information of the degradation. However, these techniques fail when the underlying parameters cannot be estimated. We propose a method to deal with situations when the underlying parameters are not known. It is based on the consensus achieved by using a set of aggregation functions and a penalty function. The method is tested in the case of a nonstationary Gaussian noise, and the Wiener filter is used to prove this methodology. The results show that the approach is consistent and it achieves comparable results for known parameters.