Improving the performance of k-means for color quantization

Abstract

Color quantization is an important operation with many applications in graphics and image processing. Most quantization methods are essentially based on data clustering algorithms. However, despite its popularity as a general purpose clustering algorithm, k-means has not received much respect in the color quantization literature because of its high computational requirements and sensitivity to initialization. In this paper, we investigate the performance of k-means as a color quantizer. We implement fast and exact variants of k-means with several initialization schemes and then compare the resulting quantizers to some of the most popular quantizers in the literature. Experiments on a diverse set of images demonstrate that an efficient implementation of k-means with an appropriate initialization strategy can in fact serve as a very effective color quantizer.

Efficient, Edge-Aware, Combined Color Quantization and Dithering

“…The original approaches mainly utilized clustering in color space. Representative clustering approaches have been based on median-cut [2], octrees [3], self-organizing maps [4], minmax [10], k-means [1], fuzzy c-means [11], adaptive distributing units [12], and variance-cut based on Lloyd-Max iterations [13].…”

“…A color quantization process typically involves two steps: color palette construction, and pixel map assignment [1]. Since natural images may contain hundreds of thousands of different colors, obtaining a visually pleasing quantized representation using a small color palette is a difficult problem.…”

“…Some of these methods use specific data structures to divide the color space, such as median-cut [2], or octrees [3]. Others use iterative clustering methods, such as k-means [1] self-organizing maps [4], to organize all colors into a limited number of clusters. All colors in the same cluster are then represented using the same color, resulting in a quantized representation.…”

On a predictive scheme for colour image quantization

“…In the second category, several clustering−based algo− rithms have been proposed [6][7][8][9][10][11][12][13]. The K−means algorithm is the well−known clustering algorithm for data grouping [6].…”

“…The K−means algorithm is the well−known clustering algorithm for data grouping [6]. Some modified algorithms such as the fuzzy c−means algorithm [7][8], the genetic c−mean algorithm [9], the co− lour finite−state LBG algorithm [11] and the improved k−means algorithm [12] have been proposed. In general, the clustering−based algorithms consume more computational cost than those of the splitting−based algorithms.…”