An Improved Median-Cut Algorithm of Color Image Quantization

The main aim of color quantization is to find the best color palette with the least differences between the original image and the quantized one. The study described in this paper proposes an improved median-cut algorithm which, by improving the pre-quantization precision, calculating the cutting position based on variance and searching reversely the colormap, significantly promotes both the speed and quality of the color quantization. The effectiveness of this algorithm has been proved by experiments.