Local Bayesian image restoration using variational methods and gamma-normal distributions

Javier Mateos*, Tom E. Bishop, Rafael Molina, Aggelos K. Katsaggelos

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Scopus citations

Abstract

In this paper we present a new Bayesian methodology for the restoration of blurred and noisy images. Bayesian methods rely on image priors that encapsulate prior image knowledge and avoid the ill-posedness of image restoration problems. We use a spatially varying image prior utilizing a Gamma-Normal hyperprior distribution on the local precision parameters. This kind of hyperprior distribution, which to our knowledge has not been used before in image restoration, allows for the incorporation of information on local as well as global image variability, models correlation of the local precision parameters and is a conjugate hyperprior to the image model used in the paper. The proposed restoration technique is compared with other image restoration approaches, demonstrating its improved performance.

Original languageEnglish (US)
Title of host publication2009 IEEE International Conference on Image Processing, ICIP 2009 - Proceedings
PublisherIEEE Computer Society
Pages129-132
Number of pages4
ISBN (Print)9781424456543
DOIs
StatePublished - 2009
Event2009 IEEE International Conference on Image Processing, ICIP 2009 - Cairo, Egypt
Duration: Nov 7 2009Nov 10 2009

Publication series

NameProceedings - International Conference on Image Processing, ICIP
ISSN (Print)1522-4880

Other

Other2009 IEEE International Conference on Image Processing, ICIP 2009
Country/TerritoryEgypt
CityCairo
Period11/7/0911/10/09

Keywords

  • Bayes procedures
  • Gamma-normal distributions
  • Image restoration
  • Variational methods

ASJC Scopus subject areas

  • Software
  • Computer Vision and Pattern Recognition
  • Signal Processing

Fingerprint

Dive into the research topics of 'Local Bayesian image restoration using variational methods and gamma-normal distributions'. Together they form a unique fingerprint.

Cite this