Accelerated pixel mapping scheme for colour image quantisation

Abstract

A fast pixel mapping scheme that is equivalent to the full search algorithm for colour image quantisation is proposed in this paper. The palette is first sorted by the mean values of colours before the pixel mapping procedure is executed. A filter that exploits the relationship between the squared sum distance and the squared Euclidean distance is employed to speed up the closest colour search process. In addition, another filter based on the triangle inequality is employed to further cut down the computational cost of the pixel mapping procedure. According to the experimental results, the proposed algorithm requires less computational cost than these comparative schemes.