语音信号中的情感信息处理研究 毕业论文+任务书+开题报告+文献综述+外文翻译及原文+MATLAB代码
语音信号中情感信息处理研究
【摘要】随着互联网的发展，语音交互技术，包括自动语音识别、合成语音和自然语言处理，开始对商业和个人电脑的使用产生重大影响。语音信号和面部表情一样，可以传达情感信息。言语情感的研究对智能人机交互具有重要的现实意义。本文介绍了语音情感识别的研究背景和相关技术，重点介绍了国内外语音情感处理的现状和发展方向。本文对语音情感识别的分析过程和设计思想进行了深入的探讨。本文完成语音信号预处理、汉明窗、MFCC、ZCPA等操作提取情感特征参数，并利用神经网络构成分类器来做情感的识别。进行很多的实验，提取出各种情绪特征参数，这些参数对不同的情绪有不同的贡献。然后利用提取的特征参数构造分类器，该分类器采用神经网络模型。最后，验证了模型的有效性。
【关键词】语音交互，语音情感识别，汉明窗，MFCC，ZCPA，神经网络
・
Research on emotion information processing
in speech signal
・
【Abstract】With the development of the Internet, voice interaction technologies, including automatic speech recognition, synthetic speech and natural language processing, are beginning to have a significant impact on the use of business and personal computers. Voice signals, like facial expressions, can convey emotional information.The study of speech emotion has important practical significance for intelligent human-computer interaction.This paper introduces the research background and related technologies of speech emotion recognition, and focuses on the current situation and development direction of speech emotion processing at home and abroad.In this paper, the analysis process and design idea of speech emotion recognition are deeply discussed.In this paper, the speech signal preprocessing, hamming window, MFCC, ZCPA and other operations are completed to extract emotional characteristic parameters, and the neural network is used to constitute a classifier to realize emotion recognition.A large number of experiments were conducted to summarize and analyze the various emotion characteristic parameters extracted, which have different contributions to different emotions.Then the extracted feature parameters are used to construct the classifier, which adopts the deep neural network model.Finally, the validity of the model is verified.
【Key Words】Voice interaction, Speech emotion recognition, Hamming window，MFCC, ZCPA, Artificial Neural Network
目 录
图目录
表目录
