PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Color Constancy Using Stage Classification
Rui Lu, Arjan Gijsenij, Theo Gevers, Koen E.A. van de Sande, Jan-Mark Geusebroek and De Xu
In: IEEE International Conference on Image Processing(2009).

Abstract

The aim of color constancy is to remove the effect of the color of the light source. Since color constancy is inherently an ill-posed problem, different assumptions have been proposed. Because existing color constancy algorithms are based on specific assumptions, none of them can be considered as universal. Therefore, how to select a proper algorithm for a given imaging configuration is an important question. In this paper, image stage models are used to aid the selection of a specific color constancy algorithm. Image stages are 3D models of a scene. Based on stage classification, the most suitable color constancy algorithms is selected. Experiments on large scale image datasets show that the proposed algorithm using stage classification outperforms state-of-the-art single color constancy algorithms with an improvement of almost 8%.

EPrint Type:Conference or Workshop Item (Paper)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Machine Vision
ID Code:6191
Deposited By:Christof Monz
Deposited On:08 March 2010