首先分析并指出了对于零数编码来说，小波变换时应当使用原始滤波器系数参与运算，而对滤波器系数进行缩放会影响其编码性能，此外，低通滤波器时延参数对零树编码性能也有重要影响，为了保证小波变换后各小波子带图象之间的自相似性，就必须使低通滤波器的时延为零，在此基础上，针对具有二进系数的双正交小波滤波器，提出了一种适合零树编码的快速小波变换算法，此算法不但以移位代替了乘法运算，而且与使用原始滤波器系数的变换等效，最后，结合快速零树编码算法，实现了一种快速高效的编码系统。

This paper points out that for the embedded zerotree wavelet(EZW) image coding, the discrete wavelet transfor(DWT) must be calculated with the original coefficient of the filters,scaling the coefficient of the filters can influence the performance of the coding.The delay of the low pass filter also has important effect on the coding performance,in order to ensure the self similarity among the images of the wavelet subbands,the delay of the low pass filter must be zero.By using the character of biorthogonal wavelets,which have dyadic rational filter coefficients,this paper presents a fast multiplierless DWT algorithm,which is very suitable for EZW image coding.The multiplierless DWT can be performed with only bit shift and integer addition operations,and moreover,the result of the multiplierless DWT is equivalent to the original DWT which is calculated with the original coefficient of the filters.At last,combining with the fast EZW algorithms,a fast and high performance image coding system is presented.