北方交通大学
纸质出版:1990
移动端阅览
[1]俞一彪,袁保宗.连续语音识别中句法结构知识的利用[J].电子学报,1990(06):68-74.
Yu Yibiao, Yuan Baozong. Use of Syntactic Structure Knowledge for Continuous Speech Recognition[J]. Acta Electronica Sinica, 1990, (6): 68-74.
本文论述连续语音识别中如何利用句法结构知识。首先介绍句法结构模型的建立
然后阐述如何基于句法结构模型分析
从语音处理的初步结果搜索出输入语句的方法。本文提出了三种启发式搜索策略:(1)最大似然优先搜索(MLF);(2)局部最大似然优先(PMLF);(3)最大似然近邻法(MLNN)。
This paper describes how to use syntactic structure knowledge for continuous speech recognition. At first
the method for constructing syntactic structure model was introduced
then describes how to search the final recognition result based on syntactic structure model and initial results from speech processing part. Three heuristic search algorithms are proposed: (1) Maximum Likelihood First (MLF) search;(2) Partial Maximum Likelihood Firs (PMLF) search;(3) Maximum Likelihood Nearest Neighbour (MLNN) search.
0
浏览量
68
下载量
4
CSCD
关联资源
相关文章
相关作者
相关机构
京公网安备11010802024621