Generalized Least Squares-Based Parametric Motion Estimation Under Non-uniform Illumination Changes

Authors:
Raúl Montoliu;Filiberto Pla
Affiliations:
Dept. Arquitectura y Ciéncias de los Computadores, Jaume I University, Castellón, Spain 12071;Dept. Lenguajes y Sistemas Informáticos, Jaume I University, Castellón, Spain 12071
Venue:
ICIAR '08 Proceedings of the 5th international conference on Image Analysis and Recognition
Year:
2008

Citing 8
Cited 0

Spline-Based Image Registration

International Journal of Computer Vision
Revised Definition of Optical Flow: Integration of Radiometric and Geometric Cues for Dynamic Scene Analysis

IEEE Transactions on Pattern Analysis and Machine Intelligence
MLESAC: a new robust estimator with application to estimating image geometry

Computer Vision and Image Understanding - Special issue on robusst statistical techniques in image understanding
Color Invariance

IEEE Transactions on Pattern Analysis and Machine Intelligence
Removing Shadows from Images

ECCV '02 Proceedings of the 7th European Conference on Computer Vision-Part IV
Distinctive Image Features from Scale-Invariant Keypoints

International Journal of Computer Vision
Robust motion estimation under varying illumination

Image and Vision Computing
Illumination intensity, object geometry and highlights invariance in multispectral imaging

IbPRIA'05 Proceedings of the Second Iberian conference on Pattern Recognition and Image Analysis - Volume Part I

Quantified Score

Hi-index	0.00

Visualization

Abstract

The estimation of parametric global motion has had a significant attention during the last two decades, but despite the great efforts invested, there are still open issues. One of the most important ones is related to the ability to recover large deformation between images in the presence of illumination changes while kipping accurate estimates. In this paper, a Generalized least squared-based motion estimator is used in combination with a dynamic image model where the illumination factors are functions of the localization (x,y) instead of constants, allowing for a more general and accurate image model. Experiments using challenging images have been performed showing that the combination of both techniques is feasible and provides accurate estimates of the motion parameters even in the presence of strong illumination changes between the images.