论文部分内容阅读
本文提出了基于改进禁止搜索 (TS)算法的矢量量化 (VQ)码书设计方法 .禁止搜索算法的关键是如何定义一个解以及如何在当前解的基础上生成邻域解 .由于码书设计的两个优化准则是最邻近条件和聚类质心条件 ,本文提出了两种禁止搜索算法的解描述方案 ,其相应算法分别叫基于码书的禁止搜索 (CB TS)算法和基于聚类划分的禁止搜索 (PB TS)算法 .为了提高禁止搜索算法的性能 ,文中在禁止搜索算法中融入了模拟退火 (SA)机制 .为了进一步提高码书性能 ,文中还将码书设计的传统LBG算法融入禁止搜索算法中 .结果表明 ,基于禁止搜索的两种码书设计方案所生成的码书性能都比LBG算法有明显提高 .
In this paper, vector quantization (VQ) codebook design method based on the improved no-to-search (TS) algorithm is proposed. The key to prohibiting the search algorithm is how to define a solution and how to generate the neighborhood solution based on the current solution. Two optimization criteria are the nearest neighbor condition and the clustering centroid condition. In this paper, two solution schemes of barring search algorithm are proposed. The corresponding algorithms are called codebook-based barring search (CB TS) algorithm and clustering-based barring Search (PB TS) algorithm.In order to improve the performance of the forbidden search algorithm, a simulated annealing (SA) mechanism is incorporated in the forbidden search algorithm.In order to further improve the performance of the codebook, the traditional LBG algorithm of the codebook design is further integrated into a forbidden search Algorithm.The results show that the performance of the codebook generated by the two codebook designs based on no search is significantly higher than that of the LBG algorithm.