无损图像编码中基于SPIHT的一种改进算法

An improved algorithm based on SPIHT in lossless image coding

为了尽可能地增加图像无损压缩编码的效率,在嵌入零树小波(EZW)算法和分层树集合分割排序(SPIHT)算法的基础上,增加了一个对A类型直接后继子孙的检测,使得重要父系数的产生依赖于非直接后继节点,并增加一C类型集来对不重要直接子孙集进行单独的分类。当检测到重要集合的数目明显大于不重要集合的数目时,对A,B或C类型的检测就可以省去。实验结果证明,无论是灰度图像或者是彩色图像,改进的算法都比传统的SPIHT算法有明显的优越性。

Aiming at increasing the efficient of image coding as much as possible,based on embedded zerotree wavelet（EZW） coding algorithm and set partitioning in hierarchical trees（SPIHT） algorithm,a new test on direct descendants of type A was added to process the parent coefficients that were significant due to their nondirect descendants,and some new sets of type C were defined to perform a separate sorting of the sets that had insignificant children.When the number of significant sets were much higher than that of insignificant,all tests over the entries（A,B and C） could be removed.Experiments show that the improved algorithm is more superior than the conventional SPIHT in both greyscale and colour images.