Iterative evaluation of regularization parameters in regularized image restoration with wavelet filter banks

Stephanakis, I; Doulamis, N; Doulamis, A; Kollias, S

dc.contributor.author	Stephanakis, I	en
dc.contributor.author	Doulamis, N	en
dc.contributor.author	Doulamis, A	en
dc.contributor.author	Kollias, S	en
dc.date.accessioned	2014-03-01T01:48:36Z
dc.date.available	2014-03-01T01:48:36Z
dc.date.issued	1999	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/25531
dc.relation.uri	http://www.scopus.com/inward/record.url?eid=2-s2.0-4944251691&partnerID=40&md5=3a2c4384d7a34c51692a48165f4b6c1f	en
dc.subject	Filter banks	en
dc.subject	Iterative methods	en
dc.subject	Multiresolution	en
dc.subject	Regularized image restoration	en
dc.subject	Wavelets	en
dc.subject.other	Image quality	en
dc.subject.other	Iterative methods	en
dc.subject.other	Matrix algebra	en
dc.subject.other	Parameter estimation	en
dc.subject.other	Signal to noise ratio	en
dc.subject.other	Spurious signal noise	en
dc.subject.other	Wave filters	en
dc.subject.other	Filter banks	en
dc.subject.other	Multiresolution	en
dc.subject.other	Regularized image restoration	en
dc.subject.other	Wavelets	en
dc.subject.other	Image retrieval	en
dc.title	Iterative evaluation of regularization parameters in regularized image restoration with wavelet filter banks	en
heal.type	journalArticle	en
heal.publicationDate	1999	en
heal.abstract	A novel approach of iterative regularized restoration of images based upon multirate representations is proposed in this paper. Regularized restoration is a method of solving the ill-posed inversion problem of image restoration. Inversion of usual degradations without regularization enhances the noise in the degraded image and may not be employed in practice. Wavelet filter-banks designed upon arbitrarily sampling lattices are proposed in order to replace the conventional regularization operator (usually a Laplacian filter) in regularized restoration of images. The proposed method employs a regularization parameter for each of the decomposition filters in the wavelet filter-bank. Thus it differs from standard regularized restoration methods which define just one regularization parameter corresponding to the smoothing filter. The regularization parameters should be estimated in advance or iteratively. A good estimate guaranties a good quality of the restored image. Statistical techniques like Generalized-Cross-Validation (GCV) may be used for estimating the regularization parameters in advance whereas the current estimate of the restored image may be used in estimating the regularization parameters in an iteratively solution of the regularization equation. A perfect reconstruction filter-bank can be used to represent the degradation filter. Factorizations of unitary matrices using Givens rotations allow for efficient representations for a variety of degradations. Should both the degradation and the smoothing filter be replaced by multirate systems, the restoration problem may be split into independent restoration problems in-each transformation channel. Regularization parameters are evaluated iteratively in each channel using image information from the corresponding subband and other channel dependent parameters. Numerical results indicate better ISNR (Improvement in Signal-to-Noise-Ratio) figures than conventional iterative regularization methods.	en
heal.publisher	World Scientific and Engineering Academy and Society	en
heal.journalName	Computational Intelligence and Applications	en
dc.identifier.spage	348	en
dc.identifier.epage	353	en