目录 摘要……………………………………………………………………(1) ABSTRACT………………………………………………………………(1) 第1章引言…………………………………………………………(2) 本课题研究背景…………………………………………(2) 方案简介……………………………………………………(2) 第2章语音信号识别的理论基础……………………………………(3) ………………………………(3) ………………………………(4) ………………………………………(5) ……………………………………(7) 第3章语音疲劳度的特征参数提取方案…………………………(9) ……………………………………………(9) ……………………………………………(10) 第4章概率神经网络………………………………………………(11) 第5章实验方案及讨论结果…………………………………………(13) ……………………………………(13) ……………………………………………………(13) ……………………………………………(14) 第6章总结与展望……………………………………………………(16) 附录(主要程序) ………………………………………………………(18) 参考文献………………………………………………………………(21) 致谢……………………………………………………………………(22) 基于语音的疲劳度检测算法研究 摘要 疲劳是一种自然现象,是人体的一种自我调节和保护功能。检测疲劳状态对于当今社会从事各行各业都有积极意义。本课题提出了一种基于语音特征参数和概率神经网络的语音疲劳度识别模型。通过训练不同时段的语音样本来构成语音源库,并建立综合识别系统。实验结果表明本方法能够反应其当时的疲劳程度,参数融入了人耳的听觉特性,故从测试结果来看,参数。 关键词:语音、疲劳度、线性预测倒谱系数、梅尔频率倒谱系数、概率神经网络 Research of Detecting Fatigue Arithmetic in Speech ABSTRACT Fatigue is a natural phenomenon which is the human body a kind of self-regulation and protection. Detection of fatigue states has positive significance in all occupation in today's society. This issue presents a feature-based parameters and the probabilistic work speech recognition model to detecting fatigue. Through training at different times of voice samples to form the voice source and to establish prehensive identification system. Experimental results show that this method can reflect its degree of fatigue at the time, parameters of the human ear into the auditory characteristics, and therefore the results from the test point of view, it's better than the parameters. KEYWODRS: Speech、Fatique、、、PNN 第一章引言