Journals Proceedings

Abstract

This paper introduces a hybrid computational technique that combines a vector to scalar mapping and single key numerical sorting to extract the color palette of an image coded in RGB (Red-Green-Blue) space. Thus a data set of color pixels in three dimensional integer space is mapped to an appropriate scalar data set that can be used to perform a fast numerical sort. The resulting ordered data set can be readily scanned to count all distinct colors and then remapped to its original dimensionality to obtain the image color palette. Explicit direct and inverse mappings are provided including several examples to illustrate the proposed mapping-sorting technique.