Registration of joint geometric and radiometric image deformations

Shahar Z. Kovalsky, Rami Hagege, Joseph M. Francos

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

2 Scopus citations

Abstract

We consider the problem of object registration where the observation and template differ both geometrically and radiometrically. The type of radiometric deformations we consider falls into the general framework of color constancy, and the geometric deformations are affine. Since the registration procedure employs the intensity information in both the template and observation, a solution to the color constancy problemhave to be established jointly with a solution to the problem of estimating the geometric transformation. We show that the original high-dimensional nonconvex search problem, that needs to be solved in order to register the observation to the template, can be replaced by an equivalent problem, expressed in terms of a sequence of two linear systems of equations. A solution to this sequence provides an exact solution to the registration problem.

Original languageEnglish
Title of host publicationProceedings of the 9th IASTED International Conference on Signal and Image Processing, SIP 2007
Pages496-499
Number of pages4
StatePublished - 1 Dec 2007
Event9th IASTED International Conference on Signal and Image Processing, SIP 2007 - Honolulu, HI, United States
Duration: 20 Aug 200722 Aug 2007

Publication series

NameProceedings of the 9th IASTED International Conference on Signal and Image Processing, SIP 2007

Conference

Conference9th IASTED International Conference on Signal and Image Processing, SIP 2007
Country/TerritoryUnited States
CityHonolulu, HI
Period20/08/0722/08/07

Keywords

  • Computer vision
  • Image registration
  • Multidimensional signal processing
  • Nonlinear parameter estimation

Fingerprint

Dive into the research topics of 'Registration of joint geometric and radiometric image deformations'. Together they form a unique fingerprint.

Cite this