| 一种基于后验概率差值的拒识算法 |
| |
| 引用本文： | 李莹莹,王成友,蔡宣平.一种基于后验概率差值的拒识算法[J].应用声学,2004,23(5):32-35. |
| |
| 作者姓名： | 李莹莹 王成友 蔡宣平 |
| |
| 作者单位： | 国防科学技术大学电子科学与工程学院,长沙,410073 |
| |
| 摘 要： | 本文提出了一种孤立词语音识别系统中基于后验概率差值的拒识算法。研究了作为拒识特征时，输入词的后验概率和后验概率差值之间的区别，并将多层感知人工神经网络用于拒识特征的学习。相比现存的几种拒识算法，本算法几乎不需要额外的计算和存储量。当识别率为98．2％时，拒识率达到了95．4％。
|
| 关 键 词： | 语音识别 拒识算法 后验概率差值 多层感知人工神经网络 后验概率 |
An out-of-vocabulary rejection based on likelihood difference for speech recognition |
| |
| LI Ying-Ying,WANG Cheng-You and CAI Xuan-Ping.An out-of-vocabulary rejection based on likelihood difference for speech recognition[J].Applied Acoustics,2004,23(5):32-35. |
| |
| Authors: | LI Ying-Ying WANG Cheng-You and CAI Xuan-Ping |
| |
| Institution: | College of Electronic Science and Engineering, National Univ. of Defense Technology, Changsha 410073;College of Electronic Science and Engineering, National Univ. of Defense Technology, Changsha 410073;College of Electronic Science and Engineering, National Univ. of Defense Technology, Changsha 410073 |
| |
| Abstract: | This paper proposes an algorithm based on likelihood difference for isolated-word speech recognition system. We explore the behaviors of likelihood and likelihood difference as discriminative features for Out-Of-Vocabulary (OOV) rejection. The Multi Layer Perceptron (MLP) is used to map the features to the OOV rejection. Compared with the existent methods, the additional amount of computation and storage is negligible in the presented algorithm, and the rejection rate can be 95.4% with an acception rate of 98.2%. |
| |
| Keywords: | Speech recognition Rejection Likelihood difference Multi-layer perceptron |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《应用声学》浏览原始摘要信息 |
| 点击此处可从《应用声学》下载免费的PDF全文 |
