提出一种基于多小波变换的改进的彩色图像分层树集合分裂（set partitioning in hierarchical trees, SPIHT）算法，将彩色RGB图像转换到YCbCr色彩域，Y通道分配到2倍于Cb,Cr的比特，在各色彩通道间构造新的方向树结构，重组图像多小波分解系数，进行嵌入式多小波彩色图像SPIHT编码.结果表明，该算法具有良好的编码效果，性能优于9/7单小波编码.

We proposes an improved color set partitioning in hierarchical trees (SPIHT) algorithm based on multiwavelet. The algorithm converts a RGB color image into the YCbCr format. Clearly, bits assigned to luminance Y plane are twice than that Cb and Cr chrominance planes. A new spatial orientation tree structure is developed between channel planes based on SPIHT coding. We embed multiwavelet SPIHT codes for color images with the re-combined multiwavelet coefficients. The results show that performance of the proposed coding method is superior to SPIHT on 9/7 wavelet.