快速分形图象压缩算法  

A Fast Algorithm of Fractal Image Compression

作　　者：蔡芳[1] 徐乃平[1] 孙隆和 

机构地区：[1]西北工业大学 [2]中航总613所

出　　处：《西北工业大学学报》1999年第3期355-359,共5页Journal of Northwestern Polytechnical University

摘　　要：分形块编码法能提供高压缩比、高质量的图象，其美中不足之处在于搜索最佳匹配块时耗时过长。本文提出的快速算法，采用二叉树和链表结构进行搜索，大大提高了匹配速度，从而缩短了编码时间。本算法分两次使用分形块编码，第二次是对第一次编码还原图和原图之间的误差进行编码，减少了图象失真。仿真结果表明，本算法较传统分形块编码法速度提高了几十倍。在高压缩比下，图象质量高于 Ｊ Ｐ Ｅ Ｇ算法。Fractal block coding which offers high compression ratio and high quality images has promising future in the field of digital image compression. However, its shortcoming is that the search time of the domain block pool must be long in the coding process. We propose a new fast encoding algorithm which arranges the domain blocks in the binary tree and chain structure so that it can shorten the encoding time and speed up the coding process by a factor of several dozens. In this proposed algorithm, fractal block encoding process is performed twice. The error image between an original image and the first reconstructed image is coded in the second encoding process. The simulation shows that the quality of reconstructed image of proposed algorthm is higher than that of JPEG at high compression ratio.

关 键 词：分形块编码 迭代函数系统 图象压缩 图像处理 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TN919.8[自动化与计算机技术—计算机科学与技术]