基于语音情绪识别方法的电路设计与仿真  

Circuit Design and Simulation Based on Speech Emotion Recognition Method

作　　者：郭颖奇[1] 栗悦 刘新怡 孟雨璇 何凌[1] 莫思特[1] GUO Yingqi;LI Yue;LIU Xinyi;MENG Yuxuan;HE Ling;MO Site(School of Electrical Engineering and Information,Sichuan University,Chengdu 610065,China)

机构地区：[1]四川大学电气信息学院

出　　处：《实验科学与技术》2019年第6期137-142,共6页Experiment Science and Technology

基　　金：国家自然基金青年科学基金(61503264);2018年四川大学大学生创新创业项目(C2018102101)

摘　　要：利用语音特征元素自动识别人的情感状态已成为智能医疗的研究热点。基于单片机技术和语音声学特征原理,针对语速、幅值、基音频率等特性进行综合分析,设计了一种针对双相情感障碍的语音情绪识别电路。该电路包含信号采集、阈值比较和计数计算等6个核心模块,实现了语音信号的有效捕捉、分析和情绪状态的二分类判别。通过采用实际语音样本和信号模拟,利用Multisim软件对电路进行仿真分析,证明了该电路的实效性和正确性。该电路设计结构紧凑,有较好的针对性和验证性,具有一定教学价值。The use of voice feature elements to automatically recognize people's emotional state has become a research hotspot of smart medical care.Based on the principle of MCU technology and speech acoustics,this paper analyzes the characteristics of speech rate,amplitude and fundamental frequency,and designs a speech emotion recognition circuit for bipolar disorder.The circuit includes six core modules,such as signal acquisition,threshold comparison and counting calculation,which realizes the effective capture,analysis and two-class discrimination of the emotional state of the speech signal.By using actual speech samples and signal simula⁃tion,the simulation of the circuit is carried out by Multisim software,which proves the effectiveness and correctness of the circuit.The circuit design is compact,has good pertinence and verification,and has certain teaching value.

关 键 词：语音特征 单片机技术 情绪识别 二分类 

分 类 号：R318.6[医药卫生—生物医学工程] TM930.2[医药卫生—基础医学]