摘要
概述语音识别技术的基本原理,对当前三种主要识别技术——动态时间规整技术、隐含马尔科夫模型技术及人工神经网络技术进行比较,重点介绍基于离散隐马尔科夫模型(DHMM)的语音识别系统的实现。
The condition and the basic principle of speech recognition technology are introduced, three different kinds of speech recognition systems such as DTW, HMM, ASR are compared, and place emphasis on how to realize DHMM in speech recognition system is presented emphatically.
出处
《河北省科学院学报》
CAS
2007年第2期8-11,共4页
Journal of The Hebei Academy of Sciences
关键词
语音识别
隐马尔科夫模型
动态时间规整
人工神经网络
Speech recognition
Hidden Markov Model ( HMM )
Dynamic Time Warping ( DTW )
Artificial Neural Network (ANN)