基于子块显著不相关检验的快速分形图象编码

A Fast Fractal Image Coding Based on Tests of Significant Uncorrelation of Blocks

图象分形编码中，子块搜索匹配的实质是子块间的相关匹配。基于这一认识，针对分形编码计算量过大、编码时间过长的缺点，本文提出先对子块进行显著不相关检验，再进行子块匹配的分形编码方法。其编码速度比基于子块分类的分形编码方法快５～２０倍（对测试图象Ｌｅｎｎａ２５６×２５６×８ｐｐｂ而言），而解码图象质量（ＰＳＮＲ）没有明显下降，压缩比还有一定的提高。

Proposes a fast fractal image coding method, which is based on tests of significant uncorrelation of blocks. For the standard test image (Lenna 256×256×8 ppb), comparing with the method of classification of blocks, the authors get a record of acceleration factors from 5 up to 20 with minor degradation in PSNR and some improvement of the compression ratio.