摘要
分形图像因其图像块的自相似性可以得到较高的压缩比、有效的减少冗余。在分形和局部迭代函数系统理论的基础上,介绍了一种灰度排列法的编码方案。该方法按照1/4子快的灰度均值,通过放射变换将图像块分为3大主类。每个子块在相应的父块类中去寻找最佳匹配的父块,从而使无效父块在搜索过程中的比较次数明显减少。实验结果表明,该算法能有效的压缩图像,压缩比高、解码图像质量好、压缩时间较短。
Fractal image compression relies on the assumption that image redundancy is efficiently captured through piecewise self-similarity on a block-wise basis. An improved global search algorithm is introduced using the gray-scale arrangement method based on the theory of fractal and local iterative function system. According to the average gray-scale of the 1/4 sub_blocks, the blocks are classified into three main kinds through affine transformation. Child-blocks found the best matching father-blocks in the corresponding kind, so that the number of times of comparison is distinctly reduced in the search for ineffective father-block. Both theoretical analyses and experiments show that high compression ratio and image quality is achieved while the compression time is comparatively reduced.
出处
《计算机工程与设计》
CSCD
北大核心
2007年第4期890-892,共3页
Computer Engineering and Design
关键词
迭代函数系统
父块
子块
分形图像
仿射变换
iterated function system
range block
domain block
fractal image
affine transformation
