摘要
在矢量量化(VQ)系统中,搜索算法是实现编码的关键技术.本文介绍非二元树搜索矢量量化系统的码本设计和搜索算法.计算机模拟实验结果表明,非二元树搜索矢量量化在失真性能和存储量方面明显优于二元树搜索矢量量化,而计算量只有全搜索矢量量化的几分之一.
The searching algorithm is a key to encoding in vector quantization(VQ)
systems.This paper describes a non-binary tree search VQ and its codebook
design algorithm and searching algorithm.The experimental results show that
the distortion performance of non-binary tree search VQ is better than that
of binary tree search VQ While its storage requirment is lower than the
latter's,and that the computation for non-binary tree search is only a frac-
tion of that for full search VQ.