汉语语音识别的关键词和集外词检测的研究与应用-模式识别与智能系统专业论文.docx

摘 摘 要 语音识别技术是人类迈向高度智能化和自动化的信怠社会所必备的 关键技术之一。经过几十年的艰苦探索和研究，语音识别研究获得了极 大的发展，某些比较成熟的技术已经逐步应用于日常生活中。以命令词 识别、关键词检测和连续数字串识别为代表的中小词汇量语音识别技 术，是语音识别实用化研究中相当重要的方向。 本文开展的工作主要集中于汉语集外调检测和关键词检测技术的研 究和应用，概括起来有以下几个方面： 1．针对语音识别中由于集外词的存在而可能造成的性能下降，结 合汉语命令词识别引擎，提出了几种不同的集外词检测方法，同时对这 几种方法进行了比较． 2．对关键词检测技术做了初步的探讨，开发了一个汉语关键词检测 的基本系统，实现了基于音节的填充模型和词树上的连续语音 ]／iterbi-Beam搜索。 3．结合在研究生阶段参与的一些语音识别的开发，介绍了集外词检 测和关键词检测在实际中的应用，并分析了这些应用系统面临的实际问 题和解决方案． 关键词：语音识别，集外词检测，关键词检测，命令词识别 ABSTRACTSpeech ABSTRACT Speech recognition is one of the most indispensable technologies for human to realize a highly intelligentized and fully roboticized information society in the future．With many researchers’enormous efforts，the past tens of years has witnessed significant progress in speech recognition technologies and Dart of them have already been applied in people’S daily life．The technologies of small or medium vocabulary speech recognition，such as voice command recognition，keyword spotting and continuous digit string speech recognition，are of great importance in the application’oriented study of speech recognition．。 IⅡthis Daper we focus our research on Mandarin OOV spotting and keyword spotting．There are several points in my work： 1．Because the existence of OOV words makes bad effect on the Derformanee of speech recognizer，we studied different OOV rejection approaches to reject OOV word．Meanwhile，we compare the performance of these approaches． 2．Investigate some aspects of keyword spotting technology and build a basic systern of Mandarin keyword spotting．We use atonal syllables as filler models for non—keywords and implement frame synchronous Viterbi—Beam search based on lexical tree． 3．Make introductions to the speech recognition projects 1 was involved in these years and the problems we meet in the projects· Keyword：Speech Recognition，OOV Spotting，Keyword Spotting，Voice Command Recognition 独创性声明本人声明所成交的论文是我个人在导师指导下进行的研究工作及取得的研 独创性声明 本人声明所成交的论文是我个人在导师指导下进行的研究工作及取得的研 究成果。尽我所知，除了文中特别加以标注和致谢的地方外，论文中不包含其 他人已经发表或撰写过