Color Rabbit: Guiding the distance of local maximums in illumination estimation

Abstract

In this paper the Color Rabbit (CR), a new low-level statistics-based color constancy algorithm for illumination estimation is proposed and tested. Based on the Color Sparrow (CS) algorithm it combines multiple local illumination estimations found by using a new approach into a global one. The algorithm is tested on several publicly available color constancy databases and it outperforms almost all other color constancy algorithms in terms of accuracy and execution speed.