Abstract: To overcome the long encoding time of fractal image coding, this paper uses a fast fractal image encoding method based on two-norm, and improves it by using the error formula, which can derive an inequality of block’s variance. This inequality is the basis to kick out the dismatched domain block, it can reduce the computational. Experimental results show that, compared with the two-norm method, the proposed algorithm improves the encoding speed, while provides the same image quality.

Key words: fractal coding, two-norm, image quality

摘要： 为解决分形图像编码过程中编码时间过长的问题，对基于2-范数匹配的快速分形图像编码算法进行改进，通过对误差公式的推导，得出误差与图像块方差的不等式，利用该不等式剔除不匹配的码本块，减少计算量。实验结果表明，与原算法相比，改进算法在保证解码图像质量的前提下，进一步提高了编码速度。

关键词: 分形编码, 2-范数, 图像质量

CLC Number: 

• TN919