基于HTK的孤立词语音识别  

HTK based speech recognition of isolated words

作　　者：杜宇斌 赵磊[1] DU Yubin;ZHAO Lei(School of Computer Science and Technology,Shandong University of Technology,Zibo 255049,China)

机构地区：[1]山东理工大学计算机科学与技术学院

出　　处：《山东理工大学学报（自然科学版）》2019年第5期63-69,共7页Journal of Shandong University of Technology:Natural Science Edition

摘　　要：设计孤立词的语音识别系统,对小词汇量、非特定人的语音进行识别。利用HTK(Hidden Markov Model Toolkit,隐马尔科夫模型工具箱)语音识别工具包进行系统的搭建,从语音识别的原理出发,对每一个单词建立和训练隐马尔科夫模型,探讨语音识别的基本流程和实现方法,为连续语音识别研究打下基础。实验结果显示,隐马尔科夫模型对孤立词具有良好的识别性能。在正常说话语境下,对语料库中单词的识别率可达到80%。The speech recognition system of isolated words is designed to recognize small vocabulary and non-specific human words.The system is built with the Hidden Markov Model(HTK)speech recognition tool package to establish and train the Hidden Markov Model for each command word.The basic process and implementation method of recognition lays the foundation for the future research of continuous speech recognition.The experimental results show that HTK has good recognition performance for isolated word recognition.In a normal speech context,the recognition rate of words in the corpus can reach 80%.

关 键 词：孤立词识别 HTK 隐马尔科夫模型 模型训练 

分 类 号：TP302.2[自动化与计算机技术—计算机系统结构]