基于自适应门限四叉树的分形图像压缩新方法  

New method for fractal image compression based on adaptive threshold value quadtree

作　　者：张梁斌[1] 范申[1] 

机构地区：[1]浙江万里学院计算机与信息学院,浙江宁波315100

出　　处：《计算机工程与设计》2006年第13期2322-2323,2337,共3页Computer Engineering and Design

基　　金：国家自然科学基金项目(10301029)

摘　　要：由Fisher提出的用于分形图像压缩的固定门限四叉树方法中,子块与父块的实验门限值一般由经验来设定,这直接影响了分形图像编码的效率,门限值的设定是分形图像压缩中的技术难点之一。考虑了输入图像的特点,给出了自适应门限的计算推导过程,提出了门限与子块的方差成正比的自适应门限的分形图像压缩方法。实验表明,对同类图像该方法压缩时间短,还原图像PSNR高,提高了分形压缩编码的效率。Fisher put forward the method of fixed threshold value quadtree in fractal image compression. Experiment threshold between child bolck and father bolck is always decided by experience, which directly decides the efficiency of fractal image compression and is a difficulty technology. According to inputting image characteristics, computing deductive procession and new quadtree fractal image compression on adaptive threshold value direct ratio child Variance are presented. Experimental results show that shorter compression time and higher image PSNR with the same image are obtained, and fractal image compression efficiency is improved greatly.

关 键 词：分形 压缩 四叉树 自适应 门限 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]