提出了主辅扫描混合编码的嵌入式零树小波（EZW）改进算法。使前一阈值下大系数的辅扫描编码和解码优先进行，构成一个新的辅扫描过程；合并了主扫描和新发现的大系数的辅扫描为一个新的主扫描过程，使编码和解码更精细有效。实验结果表明，与传统的EZW算法相比，该改进算法明显提高了图像压缩的峰值信噪比。

Improved embedded zero-tree wavelet (EZW for short) algorithm of dominant and subordinate pass mix-coding is presented. The subordinate pass of big coefficients that belong to a previous threshold becomes the first step during the encoding process. The dominant pass and the subordinate pass that constituted by the new big coefficients are combined to a new dominant pass. It makes encoding and decoding effective. Experimental result shows that compared to the traditional EZW, the improved algorithm can increase the peak signal to noise radio (PSNR for short) of image compression.