基于HMM的汉语连续数字语音识别-信号与信息处理专业论文.docxVIP

基于HMM的汉语连续数字语音识别 基于HMM的汉语连续数字语音识别 摘要 摘 要 连续数字语音识别在现实中具有广泛的应用前景。在互连网、通信、 军事、国防、人机交互等方面都有着重要的应用价值。 多年来，许多学者都致力于连续数字语音识别的研究。但是，同英语 连续数字语音识别情况相比，汉语连续数字语音识别的现状同实际应用的 要求还存在一定的差距。主要难点在于，首先汉语是单音节字，音节越少 语音间的混淆程度越高，存在一些容易混淆的数字对，如“2”和“8”，并 且汉语数字中还存在一些单元音数字，如“1”、“2”和“5”，这些数字的 连续发音序列(如“55”)的识别相当困难，识别过程中很容易引起数字“误 插入”和“误删除”。 为了得到一个高性能、低运算量的非特定人连续数字语音识别系统， 本文在识别算法中做了多种分析和改进，研究了结合韵律特征信息来提高 舢连续数字语音识别性能的算法。实验证明，该算法是有效的。 系统的各算法由VC编程实现，其中包括语音信号的预处理、信号特征 参数的提取、H删模型训练、连续数字的切割算法、识别算法、以及后处理 部分，并利用汉语的韵律信息来提高识别率。本文的研究采取渐进的方式， 在研究连续数字语音识别之前，先研究了孤立数字的语音识别；在研究基 于CH删的语音识别之前，先研究了基于DHMM的语音识别。 关键词：语音识别，汉语连续数字，HMM，韵律特征信息 作 者：张静亚 指导教师：俞一彪 基于HMM的汉语连续数字语音识别Abstract 基于HMM的汉语连续数字语音识别 Abstract Connected digit speech recognition is a crucial branch of continuous speech recognition．It has attracted the attention of many researchers for a long time， because it has wide applications，such as intemet，communication，military affairs，human—computer dialogue system，etc． Mandarin connected digit speech recognition is more difficult than English digit recognition because of its own pronunciation characteristic． The ultimate goal of this thesis is to impmve the connected digit speech recognition performance based on CHMM．Towards this goal，we investigate a variety of ways of utilizing prosodic information．Experiments show that the proposed algorithm can improve the connected digit speech recognition performance． Each unit of a speech recognition system is studied in this thesis，which includes the preprocessing，the extracting of acouStic features，me training of HMM，the separating and recognition of connected speech with prosodic information，etc．The all algorithms are realized on the sottware VC++．On the other hand，we examine isolated digit speech recognition before investigate eormeeted speech recognition，and we also examine speech recogrdtion based on DHMM before investigate recognition based on CHMM． Keywords：Speech Recognition，Mand撕n connected digit，HMM，Prosodic information Written by zhang JingYa Supervised by Yu Y