Multi-camera Matching under Illumination Change Over Time

Illumination differences between disjoint cameras can have a dramatic effect on the appearance of objects, thus increasing the difficulty of multi-camera object association. Although methods to model these inter-camera illumination conditions exist, they often rely on static illumination conditions and are unable to copewith unpredictable illumination changes over time. In this paper we propose a novel method for multi-camera object association based on adapting a learned intercamera illumination mapping function to new illumination conditions over time without the need for a manual training stage using new foreground objects. Comparative experiments are carried out using challenging data taken from a disjoint camera network. The results demonstrate that the proposed method outperforms a number of existing methods given changing illumination conditions.