采用树结构矢量组合的小波图像分类矢量量化  

Classified Vector Quantiztion of Wavelet Image Using Tree Structure Vector Combination

作　　者：郑勇[1] 朱维乐[1] 

机构地区：[1]电子科技大学电子技术系,成都610054

出　　处：《信号处理》2001年第4期302-307,共6页Journal of Signal Processing

摘　　要：本文提出了采用树结构矢量组合对小波图像进行分类矢量量化的新方法。该方法充分利用了子带系数的带间和带内的相关性，分类信息占用比特数少，并采用了基于人眼视觉特性的加权均方误差准则进行矢量量化，提高了量化增益。仿真结果表明，该方法实现简单，可达到很好的压缩效果。A new method which performs classified vector quantization to wavelet image using tree structure vector combination is proposed. the method utilizes correlation between the subbands and within the subbands fully, the classified information need a little bites to express, and vector quantization adopt weighted mean square error rule that using mankind visual characteristic to improve quantization gain. Simulation states that this method can be realized easily and can compress the wavelet image efficiently.

关 键 词：小波变换 矢量量化 图像分类 矢量组合 树结构 图像编码 图像处理 

分 类 号：TN919.81[电子电信—通信与信息系统] TP391.41[电子电信—信息与通信工程]