| 基于“增加树”的语音识别搜索网格的优化设计 |
| |
| 引用本文： | 黄彬,张明新,王劲林. 基于“增加树”的语音识别搜索网格的优化设计[J]. 计算机工程与应用, 2006, 42(34): 10-12,32 |
| |
| 作者姓名： | 黄彬 张明新 王劲林 |
| |
| 作者单位： | 1. 中国科学院,声学所DSP中心,北京,100080;中国科学院,研究生院,北京,100049
2. 中国科学院,声学所DSP中心,北京,100080 |
| |
| 基金项目： | 面向21世纪教育振兴行动计划(985计划) |
| |
| 摘 要： | 基于传统的语音识别搜索网格的分析,提出了一种基于“增加树”法和上文相关声学模型方法的识别搜索网格优化,该方法对于人名查询搜索网格中姓氏的大量冗余能够起到较好的优化作用,该识别搜索网格优化设计主要针对一个人名查询应用的识别内核开发提出来的。
|
| 关 键 词： | 语音识别 网格优化 增加树 |
| 文章编号： | 1002-8331（2006）34-0010-03 |
| 收稿时间： | 2006-09-01 |
| 修稿时间： | 2006-09-01 |
Speech Recognizing Network Optimization Design Based on "Growing Tree" |
| |
| HUANG Bin,ZHANG Ming-xin,WANG Jin-lin. Speech Recognizing Network Optimization Design Based on "Growing Tree"[J]. Computer Engineering and Applications, 2006, 42(34): 10-12,32 |
| |
| Authors: | HUANG Bin ZHANG Ming-xin WANG Jin-lin |
| |
| Affiliation: | 1.DSP Center, Institute of Acoustic,Chinese Academy of Sciences, Beijing 100080, China; 2.Graduate School,Chinese Academy of Sciences,Beijing 100049,China |
| |
| Abstract: | After analysis of traditional method,a kind of recognizing network optimization based on "growing tree" method and left-context-dependent model is presented to reduce the redundant nodes and transitions in network,it is designed for the core of a name retrieval system. |
| |
| Keywords: | speech recognition network optimization growing tree |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
