Iterative Projective Gabor Method for Images Filtering

摘要

This paper describes a method for de-noising images by thresholding Gabor transforms recursively. A new localized estimation of the noise standard deviation is obtained. It is shown that the algorithm converges globally using a few number of iterations. Experimental results show a remarkable improvement compared with the wavelet based de-noising methods (SureShrink and BayesShrink).