Image enhancement based on signal subspace approach

This paper describes a block-by-block basis image enhancement algorithm which uses the signal subspace method to enhance images corrupted by uncorrelated additive noise. The enhancement is performed by eliminating the noise components in the noise subspace and estimating the clean image from the remaining components in the signal subspace.