毕业设计（论文）-基于GMM说话人识别系统软件设计精选.doc

本科生毕业论文（设计） 中文题目：基于GMM说话人识别系统软件设计 英文题目：Software design of speaker recognition system based on GMM 学生姓名： 班级： 7班 学号： 学 院：通信工程学院 专 业：信息工程 指导教师： 职称： 副教授  基于GMM说话人识别系统软件设计 学 生 姓 名： 班 级：7班 学 号： 学 院：通信工程学院 专 业：信息工程系 指 导 教 师： 摘 要 语音是实现人们之间沟通的最直接与方便的手段，而实现人与计算机之间畅通无阻的交流一直是人们努力的目标。随着信息技术的发展，计算机的应用日新月异，随着时代的前进，说话人识别技术也要求用计算机来处理。说话人识别技术是一种从语音数据中提取出能反映说话人生理和心理特征参数，并用这组特征参数来鉴别说话人身份的技术。一段语音数据中包含了太多的信息，而说话人识别的一项目标就是从语音信号中提取出反映说话人特征的参数。本文详细写出了说话人识别技术中应用得比较多的参数——梅尔倒谱特征参数（MFCC）（GMM）Mel倒谱特征参数 高斯混合模型 Mel滤波器组 软件设计ABSTRACT Voice communication between people is to achieve the most direct and convenient means, and between man and computer communication has been a smooth one goal. With the development of information technology, computer applications with each passing day, with the progress of the times, speaker recognition technology is also required to deal with the computer. Speaker recognition technology is a voice data extracted from the reflected physiological and psychological characteristics of the speaker parameters, and parameters used to identify this group speaker identification technology. A voice data contains too much information, and speaker recognition, a goal is to extract from the speech signal characteristics that reflect the parameters of the speaker. This paper written by the application of speaker recognition technology more than the parameters - Mel Cepstrum parameters (MFCC) of the extraction process. Speaker recognition is divided into two parts, training and testing, this model of speaker recognition using Gaussian mixture model is (GMM). Article details the method of Gaussian mixture model training and speaker recognition when the pattern matching. This platform design using VC + +6.0 and the preparation of the speech recognition software, the software functions to read audio files, voice, signal preprocessing, extraction of speech signals in the Mel Cepstrum parameters and Gaussian mixture model t