计算机应用

• 图形图像处理(Graphics and image processi • 上一篇    下一篇

一种基于邻域搜索的分形图像编码算法

滕姿 邓辉文 杨久俊   

  1. 西南大学计算机与信息科学学院 西南大学计算机与信息科学学院,西南大学逻辑与智能研究中心 西南大学计算机与信息科学学院
  • 收稿日期:2007-12-10 修回日期:2008-02-27 发布日期:2008-06-01 出版日期:2008-06-01
  • 通讯作者: 滕姿

Improved algorithm of fractal image coding based on neighbour-searching

Zi TENG Hui-wen DENG Jiu-jun YANG   

  • Received:2007-12-10 Revised:2008-02-27 Online:2008-06-01 Published:2008-06-01
  • Contact: Zi TENG

摘要: 分形图像编码以其高压缩比,使得它具有广泛的应用前景,但是传统的分形编码方法压缩时间长,这是因为在为每一个值域块寻找合适的定义域块时需要在码本中进行大量的搜索匹配。研究发现,和值域块最匹配的定义域块大多数在值域块的附近。根据这一结论,提出了一种基于邻域的螺旋式搜索方法,并利用图像块间的信息熵差值,缩减了搜索范围。另外对误差图像进行编码,作为解码图像的补充,提高了解码图像质量。实验表明,该方法在不降低图像质量和信噪比的前提下大大加快了编码的速度。

关键词: 分形编码, 邻域, 定义域块, 值域块,

Abstract: Fractal image coding has high compression ratio, but the traditional fractal compression encoding method has the disadvantage of long processing time, because they need a lot of searching and matching for each range block to find a suitable domain block. This study shows that the domain block matching most of the range block is in the vicinity of the range block. According to this conclusion, an improved algorithm of fractal coding based on matching in range block's neighbourhood searching was presented, and the scope of the search was reduced by the information entropy difference of image block. In addition, the error image was coded as a supplement to improve the quality of the decoded image. Experimental results show that the coding time can be reduced greatly while the quality of reconstructed image is not much influenced.

Key words: fractal coding, neighbourhood, domain block, range block, entropy