A new robust nonlinear filter for noise-smoothing and edge-enhancement

The restoration and enhancement of degraded images are of fundamental importance in image processing applications. In recent years nonlinear techniques have been extensively used to achieve these tasks. One of the most important families of nonlinear image filters is based on order statistics. They have shown to posses excellent robustness properties in presence of impulsive noise while preserving the edge information. In this work we introduce a new robust nonlinear filter based on generalized maximum likelihood reasoning and order statistics. This new filter is not only capable of attenuating the noise and preserving the details, but also has the ability to sharpen the edges. In addition, it does not require any type of threshold or weighting factor and is computationally efficient. Experimental result on real images is also provided