Color image denoising and blind deconvolution using the Beltrami operator

R. Kaftory, N. A. Sochen, Y. Y. Zeevi

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

14 Scopus citations

Abstract

A new method for the recovery of noisy and blurred color images is presented. The image is reconstructed and the blurring kernel is approximated, under the assumption of linearity and spatial invariance of the blurring kernel. It is done by combining the Beltrami operator, which was introduced as a general framework for low-level vision, with the scheme of the blind deconvolution, which was introduced for the recovery of blurred and noisy gray value images. Consequently, image and kernel edges are preserved due to the adaptive smoothing feature of this operator. The color channels are coupled by a Riemannian structure, which is defined on the color image. The functional minimization scheme is presented and results of applying it in the recovery of blurred and noisy color images are illustrated.

Original languageEnglish
Title of host publicationISPA 2003 - Proceedings of the 3rd International Symposium on Image and Signal Processing and Analysis
EditorsA. Neri, H. Babic, S. Loncaric
PublisherIEEE Computer Society
Pages1-4
Number of pages4
ISBN (Electronic)953184061X
DOIs
StatePublished - 2003
Event3rd International Symposium on Image and Signal Processing and Analysis, ISPA 2003 - Rome, Italy
Duration: 18 Sep 200320 Sep 2003

Publication series

NameInternational Symposium on Image and Signal Processing and Analysis, ISPA
Volume1
ISSN (Print)1845-5921
ISSN (Electronic)1849-2266

Conference

Conference3rd International Symposium on Image and Signal Processing and Analysis, ISPA 2003
Country/TerritoryItaly
CityRome
Period18/09/0320/09/03

Keywords

  • Additive noise
  • Atmospheric modeling
  • Color
  • Colored noise
  • Convolution
  • Deconvolution
  • Gaussian noise
  • Image reconstruction
  • Kernel
  • Noise reduction

Fingerprint

Dive into the research topics of 'Color image denoising and blind deconvolution using the Beltrami operator'. Together they form a unique fingerprint.

Cite this