Classification-based color constancy

Abstract

In this work, we investigate how illuminant estimation techniques can be improved using indoor/outdoor classification. The illuminant estimation algorithms considered are derived from the framework recently proposed by Van de Weijer and Gevers. We have designed a strategy for the selection of the most appropriate algorithm on the basis of the classification results. We have tested the proposed strategy on a subset of the widely used Funt and Ciurea dataset. Experimental results clearly demonstrate that our strategy outperforms general purpose algorithms. © 2008 Springer-Verlag Berlin Heidelberg.