基于方向树结构矢量分类的小波图像网格编码矢量量化

本文提出了采用方向树结构矢量组合并分类对小波图像进行网格编码矢量量化（ＴＣＶＱ）的新方法。该方法矢量构成结合了子带系数的方向性，充分利用了子带系数带间和带内相关性，按能量和活跃度进行两级分类，降低了类中矢量的内部离散度，对活跃和非活跃类矢量实行加权ＴＣＶＱ，利用卷积编码扩展信号空间，用维特比算法搜索最优量化序列，比使用加权 ＶＱ提高了 ０．７ｄｂ左右。该方法编码计算复杂度适中，解码简单，有较好的压缩效果。

Classified Vector Quantiztion of Wavelet Image Using Orientation Tree Structure Vector Combination

A new method which performs trellis coded vector quantization(TCVQ) to wavelet image using orientation tree structure vector combination and classification is proposed. The' method accords with the orientation of subbands coefficients to combine the vectors and exploits the correlation between the subbands and within the subbands fully. It makes two stages classification by the vector's energy and activity and reduces the inner dispersion of the classified vectors. TCVQ is performed to active and inactive vectors.it uses convolution^ coding to expand signal space and uses Viterbi algorithm to find a optimized survived quantized sequence, it has an advantage of 0.7db or so over weighted VQ. The method has modest encoding complexity with simple decoding and can achieve much better compression effect.