基于主分量分析的相关矢量量化编码算法

A Coding Algorithm Using PCA-Based Correlation Vector Quantization

在编码前,首先计算码书中所有码字在主轴上的投影值,然后按照这些投影值从小到大对码字进行排序;在编码过程中,利用邻近图像块的高度相关性和当前输入矢量在主轴上的投影值共同确定相应的码字搜索范围.实验结果表明,与传统穷尽搜索矢量量化编码法相比,虽然文中算法的编码质量略有下降,但编码速度和压缩效率都有了显著的提高.

By the improved coding algorithm proposed in the paper, before encoding, we project all the code words on the principal axis which is determined by principal component analysis （PCA）, and a sorted codebook is obtained according to the ascending order of the projection values. During the encoding stage,we use the high correlation of the adjacent image blocks and the projection value of the input vector on the principal axis to determine the searching range. The experimental results show that PCA and correlation vector quantization can accelerate the encoding speed and increase the compression rate compared with the full search vector quantization, although the encoding quality degrades a little.