Restauration d'image dégradée par un flou spatialement variant

Saima Ben Hadj 1 Laure Blanc-Féraud 1
1 ARIANA - Inverse problems in earth monitoring
CRISAM - Inria Sophia Antipolis - Méditerranée , Laboratoire I3S - SIS - Signal, Images et Systèmes
Abstract : In most of the existing image restoration techniques, the blur is assumed to be spatially invariant. However, different physical phenomena related to the optic's properties makes that degradations may change according to the image's areas. In this work, we consider a piecewise-varying PSF model with smooth transitions between areas in order to take into account blur variation in the image. We develop for this model, a convenient deconvolution method by minimizing a criterion with a total variation regularization. We rely on a fast minimization method using a domain decomposition method that was recently developed by Fornasier et al. 2009. We thus obtain an algorithm where the criterion minimization is performed in a parallel way on different areas of the image, taking into account the estimates of neighboring areas of the considered sub-image, so that the final solution is space-varying deconvolved.
Document type :
Conference papers
Liste complète des métadonnées

Cited literature [9 references]  Display  Hide  Download

https://hal.inria.fr/inria-00625519
Contributor : Mikael Carlavan <>
Submitted on : Wednesday, September 21, 2011 - 5:47:12 PM
Last modification on : Monday, November 5, 2018 - 3:52:01 PM
Document(s) archivé(s) le : Thursday, December 22, 2011 - 2:36:03 AM

File

c_benhadj126.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : inria-00625519, version 1

Collections

Citation

Saima Ben Hadj, Laure Blanc-Féraud. Restauration d'image dégradée par un flou spatialement variant. GRETSI, Sep 2011, Bordeaux, France. ⟨inria-00625519⟩

Share

Metrics

Record views

316

Files downloads

569