Restoring images with regularization in uncorrelated transform domain

Abstract

Conventional spatially adaptive regularized image restoration schemes weigh the amount of regularization according to the spatial content of an image. A better performance can be achieved by first separately decorrelating the available information about the signal under analysis into uncorrelated components and then weighting the amount of regularization performed to these components accordingly. An iterative restoration algorithm is proposed accordingly to restore images which are blurred and corrupted with color noise.