文档介绍：本科毕业论文(设计) 题目:孤立词语音识别的并行编程实现 2010 年 6月 5日青岛大学本科毕业生论文(设计) Parallel Programming of Isolated Words Speech Recognition 青岛大学本科生毕业论文(设计) 摘要语音识别是解决机器“听懂”人类语言的一项技术。首先介绍了目前国内外语音识别系统的现状、发展及研究方向, 并对语音识别系统的原理、并行化的实现做了详细分析, 在此基础上进行语音识别系统的设计,主要内容有: 提取语音、端点检测、特征值提取、训练数据、语音识别。选用 HMM 隐马尔科夫模型, 基于 VC2005 编译环境下的的多线程编程,实现算法的并行运算,提升了语音识别的效率。设计完成后在实验室进行了模拟调试, 实验结果表明:所设计的程序满足语音识别系统的基本要求。关键词语音识别孤立词多线程并行特征值断点检测 Abstract Speech recognition isa techni que to solve the machine "understand" human language. First introduces the current status of speech recognition system at home and abroad, development and research, and the principle of speech recognition systems, parallel implementation ofa detailed explanation todo in this speech recognition system based on the design of main contents are: Extract voice, endpoint detection, feature extraction, training data, voice recognition. HMM hidden Markov model selection, based on VC20 piler environment of multi-threaded programming, puting algorithms to improve the efficiency of speech recognition. pletion of the design was simulated in the lab debugging, experimental results show that: the design process to meet the basic requirements of speech recognition systems. Keywords Speech Recognition Isolated Words Multi-Threads Parallel Eigenvalue Breakpoint Detection 、目录前言......................................................................................................................... 1 第1章语音识别技术概述.................................................................................... 2 语音识别的历史背景........................................................................................................ 2 国外研究历史及现状.................................................................................................. 3 国内研究历史及现状.................................................................................................. 4 语音识别技术应用............................................................................................................ 5