基于HTK的连接数字语音识别的研究  

The Research of Continuous Digital Speech Recognition Based on HTK

作　　者：黄少龙[1] Huang Shaolong(Department of Information Engineering, Taiyuan University of Technology, Jinzhong Shanxi 030600, Chin)

机构地区：[1]太原理工大学信息工程学院,山西晋中030600

出　　处：《山西电子技术》2016年第5期86-88,共3页Shanxi Electronic Technology

基　　金：山西省留学回国择优资助项目(晋人社厅函[2013]68号);山西省自然科学基金(2014021022-6)

摘　　要：HTK是剑桥大学工程系开发的一套基于C语言的语音处理工具包,目前在语音识别、语音合成以及字符序列等领域已得到广泛应用。本文首先简单介绍了语音识别的一些基本原理,接着从HTK的主要原理和软件结构出发,阐述了基于HTK语音识别系统的搭建过程以及每部分用到的工具或函数,最终完成了7字长的连续语音识别系统,并验证了其识别率。HTK is a set of speech processing tools based on C language developed by the University of Cambridge＇s engineering department. It has been widely used in the fields of speech recognition,speech synthesis and character sequence at present. This paper first introduces the basic principle of speech recognition,and then expounds the building of speech recognition system based on HTK and the tool or function used in every part from the basic principle and the software structure of HTK. Finally,the recognition system of 7 characters long continuous speech is completed and the recognition rate is verified.

关 键 词：HTK 语音识别 HMM模型 梅尔倒谱系数MFCC 

分 类 号：TN912.34[电子电信—通信与信息系统]