Abstract: According to the disadvantage of the long time of the fractal image coding, propose the algorithm of fractal image coding combining dispersion analysis and quadree. The new algorithm uses the dispersion analysis criteria in search, the method can search the special domain blocks that they meet the conditions of dispersion, which reduce the complexity of search process and the number of invalid domain blocks, thus finish image coding. Experimental results show that this algorithm can improve coding speed by 5.87 times and 7.84 times than the algorithm of the variance and the improved quadree division, the PSNR of image decoding dosen’t change obviously.

Key words: fractal, quadtree division, dispersion analysis, variance, image coding

摘要： 针对分形图像编码时间过长的问题，提出了差量分析和四叉树结合的分形图像编码算法。在搜索码本集合中运用差量分析匹配准则，只搜索那些符合差量分析匹配准则的定义域块，从而降低了搜索过程的复杂度，完成编码。实验表明，提出的算法较方差算法和改进四叉树分割算法速度分别提高了5.87倍和7.84倍，解码图像的峰值信噪比基本不变。

关键词: 分形, 四叉树分割, 差量分析, 方差算法, 图像编码