摘要
智能手机越来越普及,为读者提供移动图书检索服务带来了很大的方便。将语音识别技术应用到移动检索中,可以减轻读者的输入难度,进一步提升读者的检索体验。文章提出一种基于语音识别技术的图书移动检索系统框架以及一种基于图书检索定制的语音识别服务优化方案。分析了该方案与基于语音输入法和基于语音识别云二种方案的优劣。在此基础上,提出了一种基于HTK、SRILM、JUICER三个开源项目来构建的具体实施方案。
Intelligent mobile phone becomes more and more popular,providing mobile book search service can bring great convenience for the readers.With the help of speech recognition technology,it is easier for the user to input search request on the mobile phone and the user with get better experience.This paper presents a framework of book retrieval system on the mobile device based on speech recognition technology.In order to improve the accuracy rate of speech recognition service,a solution is proposed.Compared with the solution based on the voice input method and the solution based on the speech recognition cloud,the advantages of the solution proposed are analyzed.At the end,an implementation scheme based on open source projects such as HTK,SRILM,JUICER is put forward.
出处
《图书馆学研究》
CSSCI
2015年第5期41-44,共4页
Research on Library Science
关键词
图书移动检索
语音识别
语言模型
声学模型
解码器
books mobile search speech recognition language model acoustic model decoder