摘要
文章提出了一种基于范数和方差相结合的码字快速搜索算法。该算法在搜索前预先计算码书中码字的范数和方差,并按范数的大小对码书中的码字进行排序;搜索时,利用有序的码书和有效的删除准则,大大减小了码字的搜索范围,从而降低了编码时间。
A fast codeword search algorithm for vector quantization which is based on the norm and variance of codewords is presented. Before every search, the norm and variance of the codewords are computed and saved in the codebook, then the codewords are listed in the ascending order of their norms. During the encoding process, the search area is decreased greatly by using the ordered codebook and the efficient elimination criteria, and the encoding time is reduced accordingly.
出处
《合肥工业大学学报(自然科学版)》
CAS
CSCD
北大核心
2006年第3期288-291,共4页
Journal of Hefei University of Technology:Natural Science
关键词
矢量量化
码字搜索
范数排序
方差不等式
vector quantization
codeword search
norm order
variance inequality