基于相关系数的快速分形图像压缩方法  

A Fast Fractal Image Compression Method Based on Correlation Coefficient

作　　者：任卓敏[1] 林贻侠[1] 

机构地区：[1]上海大学机械电子工程与自动化学院,上海200072

出　　处：《上海大学学报（自然科学版）》2002年第6期482-485,共4页Journal of Shanghai University:Natural Science Edition

摘　　要：分形图像压缩的缺点在于编码时间过长.该文提出了一种基于相关系数的匹配判别式,可大量排除不相关的Domain(区域)块,从而可降低匹配的计算量,提高编码的速度.判别式中阈值由全局的相关系数ρ和各Range(排列)块的标准差σr所确定.本方法在大幅提高编码速度的同时,能保证良好的图像质量,这得到了实验结果的验证.The main disadvantage of fractal image compression is its long coding time. This paper proposes a matching discriminant based on correlation coeffient, which excludes a large quantity of irrelevant domain blocks so that the number of matching calculations is reduced. Correlation coeffient of the image and standard deviation of each range block together define the threshold in the matching discriminant. The method lead to a significant speedup of coding and good quality of reconstructed image. This is verified by experiment results.

关 键 词：分形 图像压缩 相关系数 匹配判别式 

分 类 号：TN919.8[电子电信—通信与信息系统]