Abstract: Self-similar of graphics is an important branch in the study of image fractal, the extraction and quantification of the feature of self-similar image particularly gets the most attention. In the process of fractal image coding, the extraction and quantification of self-similar image feature has been widely studied and applied. Based on the study of image characteristics of mathematical statistics, this paper puts forward a way of classifying the image blocks, which improves fractal image compression algorithm by dividing the image blocks with similar characteristics into smaller area. Experimental results show that this method improves the speed of fractal coding and reduces the amount of calculation in a great extent while reconstructed image quality can be guaranteed.

Key words: self-similar, characteristics of mathematical statistics, fractal, image compression, Iterated Function System（IFS）

摘要： 图像自相似是图像分形研究中一个重要的研究方向，尤其是对自相似图像子块的特征提取量化问题尤为引人关注。在分形图像编码发展的过程中，图像自相似特征的提取和量化得到了广泛的研究和应用。通过对图像数理统计特征的研究，提出了一种对图像子块进行分类的方法，使拥有相似特征的图像子块能够划分到更小的区域范围内，改进了分形图像压缩算法。经过分析和实验证明，该方法在不影响重建图像质量的前提下，提高了分形编码的速度，较大程度地减少了计算量。

关键词: 图像自相似, 数理统计特征, 分形, 图像压缩, 函数迭代系统