基于自适应门限的快速分形编码方法  

Fast fractal coding method based on adaptive threshold

作　　者：李万臣 贾宝 刘雪飞[1] 张俊峰[1] 

机构地区：[1]哈尔滨工程大学信息与通信工程学院,黑龙江哈尔滨150001

出　　处：《应用科技》2007年第12期25-27,共3页Applied Science and Technology

摘　　要：利用图像子块的自身统计特性,并结合父块变换过程中方差不变的特性来减少匹配库中候选块的数目,依据匹配门限可以由子块的方差来确定的原理,提出了一种基于自适应门限的分形图像编码算法的改进算法.实验表明,该算法在保持压缩比基本不变的情况下,在还原图像PSNR方面有所提高,缩短了图像的编码时间.The statistical characteristic of image child-blocks shows that the variance keeps unchanged in the conversion process of father-blocks. With this feature, the number of candidate blocks in matching base can be reduced. Therefore an improved algorithm for fractal image coding is proposed based on adaptive threshold, in which the matching threshold can be determined by the variance of child-blocks. Experimental results indicate that the performance of PSNR of recovered images is improved while the compression ratio is approximately constant, and the coding time is shortened.

关 键 词：分形 自适应门限 子块 父块 四叉树 方差 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]