Sequential product code quantization of digital color image

Abstract

A digital color image quantization mechanism employs sequential product code vector quantization, to sequentially extract chrominance and luminance values from the vectors and quantizes chrominance and luminance features based upon a conditional distribution of these features within partitioned regions of chrominance/luminance color space. The mechanism sequentially partitions a histogram of the original digital color image in luminance, chrominance (Y,Cb,Cr) space coordinates into a plurality of sub-regions or color space cells, such that each partitioned color cell is associated with a color of the output palette through which the color composition of a reproduced color image is defined. A splitting criterion determines the sequential order of partitioning of an axis. Because of the increased sensitivity of the human visual system to contouring artifacts in regions of an image to low spatial activity, the splitting criterion along the luminance axis is scaled or weighted in inverse proportion to the average spatial activity of the luminance-chrominance region subject to be split. A map of chrominance and luminance output codes is generated for the respective pixels of the output color image in accordance with the axial splitting or quantization of the chrominance and luminance components of the histogram.