Cross-channel histogram equalisation for colour face recognition

Abstract

Changes in illumination conditions will alter the appearance of digital images that will in turn have a detrimental effect on face recognition. To overcome the problem, histogram equalisation has already been applied to grey world face recognition and extended to colour object recognition by independently processing the three colour channels. This paper furthers this work by introducing a new technique, cross-channel histogram equalisation, and reports upon its application to colour face recognition under different illumination conditions. Based on the experimental tests, our approach has been shown to outperform other efforts on histogram equalisation for normalisation. Finally we give our conclusions and discuss future work. © Springer-Verlag 2003.