基于EZW图像压缩算法改进的可行性研究

Feasibility Study for the Improved Algorithm of the EZW Image Compression

在嵌入式零树小波编码(EZW)算法的基础上,为实现快速高质量数字图像压缩性能,探讨消除树间系数节点冗余以进一步实现压缩的可能性,并给出避免多次重复后代子孙系数节点扫描的改进策略.首先,根据EZW算法中全局阈值提取的重要系数在各子带的分布特性,给出一个各子带重要系数的自适应阈值序列.然后,依据归一化能量序列熵值进一步压缩各子带的重要系数,以实现消除小波树的树间冗余.最后,采用小波树逆编码完成对图像数据的编码输出.

Based on the Embeded Zero Tree wavelet coding, in order to realize high-speed image transmission with high-quality, this paper analysed the possibility of further reducing the redundancy between trees and put forward improved strategies to reduce the frequency of offspring nodes scanning. First of all, giving an adaptive threshold sequence on the basis of significant coefficients of each sub-bands; Then, further compressing the significant coefficients of each sub-bands according to the normalized energy sequence entropy to reduce the redundancy between trees. Finally, using wavelet trees inverse coding to complete the output coding of the image data.