基于DCT系数小波重组的图像压缩算法  

An Image Compression Algorithm Based on DCT Coefficients Rearranged By Wavelet Structure

作　　者：王国华[1] 陈宏[1] 俞一彪[1] 芮贤义[1] 

机构地区：[1]苏州大学电子信息学院,苏州215021

出　　处：《信息安全与通信保密》2006年第4期100-102,共3页Information Security and Communications Privacy

摘　　要：论文提出一种与小波变换结构近似的嵌入式编码—MSESPIHT。将图像8×8子块进行DCT变换后的系数按照小波变换结构重新组合,运用均值去除法降低低频段系数的冗余度。之后,运用基于分层树集合分割排序的编码算法对图像进行压缩编码。实验表明,采用MSESPIHT处理的图像,在相同压缩比的情况下,图像质量得到改善,当码率为1.00bpp时,EZW、MSEDCT、SPIHT、MSESPIHT、峰值信噪比分别为39.55dB、39.75dB、39.97dB、40.12dB。This paper introduces an embedded coding method based on the similar wavelet coefficients structure, which is said MSESPIHT. The method rearranges coefficients of 8＊8 block DCT and reduces high correlation between low frequency coefficients by a mean-subtract method. Then, image is compressed by SPIHT coding with high PSNR at high compression ratios. For EZW,MSEDCT,SPIHT , MSESPIHT, when coding rate is 1.00 bit per pixel, PSNR partly is 39.55dB,39.75dB,39.97dB,40.12dB.

关 键 词：DCT系数 小波重组 SPIHT编码 空间方向树 

分 类 号：TN911.73[电子电信—通信与信息系统]