在研究Shapiro提出的嵌入式零树小波编码EZW算法理论的基础之上，针对其效率的不足，本文提出了一种新的嵌入式小波图像编码算法：将子带进行划分，每次扫描只针对重要子带进行，从而避免大量重复扫描，大大减少要扫描和编码的零树的数量。另外，舍弃了增益很小但代价很大的副表重排序。实验证明,该算法简单，节省码流，加速编解码过程，提高了算法的效率而没有损失编码的精度。此外，对近年来出现的一些改进算法进行了分析和研究，并与本文算法作了一定比较。

A new embedded wavelet image coding scheme is proposed to improve the efficiency of EZW which was founded by Shapiro: divide subbands into important ones and unimportant ones. To avoid the repetitive scans,it is necessary to scan only the important subbands,which will decrease the number of zerotrees needed to scan and code. Besides,the resorting of the subordinate list,which pays a lot but little benefits is given up. Experiments show the simple algorithm saves code streams,speeds up the coding and decoding course,and increases the efficiency without losing the precision of coding. Moreover,we analyse and study some improved algorithms put forward recently,and compare the algorithm in this paper with them.