Local adaptivity to variable smoothness for exemplar-based image denoising and representation - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Rapport (Rapport De Recherche) Année : 2005

Local adaptivity to variable smoothness for exemplar-based image denoising and representation

Résumé

A novel adaptive and exemplar-based approach is proposed for image restoration and representation. The method is based on a pointwise selection of small image patches of fixed size in the variable neighborhood of each pixel. The main idea is to associate with each pixel the weighted sum of data points within an adaptive neighborhood. This method is general and can be applied under the assumption that the image is a locally and fairly stationary process. In this paper, we focus on the problem of the adaptive neighborhood selection in a manner that it balances the accuracy of approximation and the stochastic error, at each spatial position. Thus, the new proposed pointwise estimator automatically adapts to the degree of underlying smoothness which is unknown with minimal a priori assumptions on the function to be recovered. Finally, we propose a practical and simple algorithm with no hidden parameter for image denoising. The method is applied to both artificially corrupted and real images and the performance is very close, and in some cases even surpasses, to that of the already published denoising methods. Also, the method is demonstrated to be valuable for applications in fluorescence microscopy.
Fichier principal
Vignette du fichier
RR-5624.pdf (3.86 Mo) Télécharger le fichier
Loading...

Dates et versions

inria-00070384 , version 1 (19-05-2006)

Identifiants

  • HAL Id : inria-00070384 , version 1

Citer

Charles Kervrann, Jérôme Boulanger. Local adaptivity to variable smoothness for exemplar-based image denoising and representation. [Research Report] RR-5624, INRIA. 2005, pp.60. ⟨inria-00070384⟩
228 Consultations
100 Téléchargements

Partager

Gmail Facebook X LinkedIn More