摘要
将人类视觉系统的特性引入对值域块、定义域块的划分以及对定义域池的搜索上来,提出了一个能显著提高编码速度的分形图像压缩算法.根据HVS特性将图像分割后,搜索空间得到了极大的缩减,并且最佳匹配块只在具有相同HVS特性的块间进行,因此能够显著降低计算的复杂性.理论和实验结果表明:与叉迹算法和经典算法相比,在保持图像质量的前提下,本文算法能够显著提高编码速度和压缩比,因而是一种有效的分形图像压缩方法.
A novel algorithm for the fraetal image impression is presented in this paper by introducing the feature of Human Visual System （HVS） to the partition of the range blocks, the partition of the domain blocks, and the search method for the domain pool. The HVS-Based algorithm can significantly improve the encoding speed. After the image is partitioned by using the feature of HVS, the search space is greatly reduced and the best matching blocks is carried out only among the blocks which have the same HVS feature. Therefore, the computation complexity is drastically reduced. Theoretical and experimental results show that when HVS-Based algorithm is compared with the cross trace algorithm and the classic algorithm, it can obtain faster encoding speed and higher compression ratio while remaining a good restructured image quality. Therefore, the algorithm presented in this paper is an effective encoding method for fractal image compression.
出处
《小型微型计算机系统》
CSCD
北大核心
2008年第10期1849-1854,共6页
Journal of Chinese Computer Systems
基金
国家高技术研究发展计划“八六三”(2006AA04Z211)资助
