摘要： 针对EZW算法变换和编码不能同时进行，因而不能实现完整意义上的子图像编码，以及该算法在全图像的小波变换未完成前，算法不能被执行，从而带来算法复杂度较高等问题，通过对EZW算法的深入分析，从算法的变换阶段入手，提出一种基于分块变换-融合滤波的改进EZW算法。实验结果证明，该改进算法在保持EZW算法主要优点的同时，能较好地实现完整意义上的子图像编码，同时可以大幅度降低计算复杂度。

关键词: 嵌入式零树小波, 子图像编码, 分块变换-融合滤波, 带峰值

Abstract: The Embedded Zero-tree Wavelets(EZW) algorithm transformation and the code can not be in progress at the same time, thus they can not realize in the complete significance sub-image coding, and this algorithm can not be carried out before the total graph alike wavelet transformation has not completed, and thus brings the problems as the algorithm is more complicated. In view of these questions, through the EZW algorithm’s thorough analysis, obtained from the algorithm transformation stage, an advanced EZW algorithm based on partition transform-merge filter is proposed. Experimental result proves that this advanced algorithm maintains the EZW algorithm principal advantage, can realize in the complete significance sub-image coding well, and simultaneously may have a lower computation complexity.

Key words: Embedded Zero-tree Wavelets(EZW), sub-image coding, partition transform-merge filter, sub-band-max value

中图分类号: 

• TN919.81