| 基于K-MEANS算法的语境相关矢量量化 |
| |
| 引用本文： | 许晓斌,丁丰,林碧琴,袁保宗.基于K-MEANS算法的语境相关矢量量化[J].自动化学报,2000,26(3):369-372. |
| |
| 作者姓名： | 许晓斌 丁丰 林碧琴 袁保宗 |
| |
| 作者单位： | 1.北方交通大学信息科学研究所,北京 |
| |
| 摘 要： | 研究用于连续语音识别的语境相关矢量量化技术.提出采用k-means(k-均值)算法
逐一地调整决策树叶子所包含的各个语境,实现对音素模型的混合密度的优化.实验结果表
明,采用k-means算法的语境相关矢量量化得到的平均分布密度比简单合并决策树叶子所得
到的平均分布密度提高4％～10％.
|
| 关 键 词： | 连续语音识别 语境相关矢量量化 k-means算法 混合密度的优化 |
| 收稿时间： | 1997-12-23 |
| 修稿时间： | 1997-12-23 |
CONTEXT-DEPENDENT VECTOR QUANTIZATION BASED ON K-MEANS ALGORITHM |
| |
| XU Xiaobin,DING Feng,LIN Biqin,YUAN Baozong.CONTEXT-DEPENDENT VECTOR QUANTIZATION BASED ON K-MEANS ALGORITHM[J].Acta Automatica Sinica,2000,26(3):369-372. |
| |
| Authors: | XU Xiaobin DING Feng LIN Biqin YUAN Baozong |
| |
| Affiliation: | 1.Institute of Information Science,Northern Jiaotong University,Beijing |
| |
| Abstract: | An approach based on k-means algorithm
to implement context-dependent vector quantization for continuous speech recognition is
presented.With the approach,individual phonetic context of each leaf in a decision tree
can be moved from one leaf to another,in order to optimize the mixture density given by
that tree.Experimental results given in this paper demonstrate that the average likelihood
given by this approach based on k-means algorithm is four to ten percent higher than that
obtained with the previous method of optimizing mixture densities by merging leaves of
decision trees. |
| |
| Keywords: | Continuous speech recognition context dependent vector quantization k-means algorithm optimization of mixture density |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《自动化学报》浏览原始摘要信息 |
|
点击此处可从《自动化学报》下载全文 |
