摘要
颜色量化是利用人眼对颜色的惰性,将原图像中不太重要的相似颜色合并为一种颜色,减少图像中的颜色,而使量化前后的图像对于人眼的认识误差最小即量化误差最小。在此从揭示八叉树颜色量化算法的优缺点开始,优化八叉树结构,进化为以二叉堆和数组索引数据结构,对最不重要颜色不断地进行退化,最终达到量化要求的颜色数量。实践证明,在相同情况下,二叉堆比八叉树颜色量化对图像的量化误差更小。
Color quantization is to use the human eye on the color of inertia, the original image is similar to the less important colors into a color, then reduce the image color, the images before and after quantization of the human eye that is, the sinai -lest error of the minimum quantization error. The octree color quantization algorithm of the advantages and disadvanta- ges of starting to abandon the algorithm octree structure,and evolution to a binary heap data structure and the array index of the most important degradation of color continue to carry out,and ultimately to achieve the volume of the number of colors required, and demonstrate the algorithm code.
出处
《现代电子技术》
2010年第6期163-167,共5页
Modern Electronics Technique
关键词
颜色量化
八叉树
二叉堆
退化算法
color quantization
octree
binary heap
degradation
