摘要
以生理信号分析为主,表情行为观察和情绪主观感受评价为辅,对多名被试的情绪进行识别。60名大学女生接受恐惧-快乐-轻松的情绪诱发,有效数据55名,对应每个情绪片段,根据信号标记以及GSR微分,截取1min的生理信号进行处理和分析,应用SPSS对各生理参数进行情绪的单因素方差分析,然后采用逐步多类判别法,提取特征参数以识别情绪。结果表明HR、HRV、R波、T波各生理参数对情绪较敏感;提取出HFP,HRmax,PNN50,LF/HF,Ratio,LFP,MeanNN7个特征参数,构建情绪判别函数Fuction1,Fuction2和Z1、Z2,Z3;轻松的判别正确率为88.0%,快乐的为92.0%,恐惧的为80.0%,总体判别正确率为86.7%。以生理信号分析为主,辅助表情行为观察和情绪主观感受报告,是一种有效的情绪识别方法,所得数据客观、准确,提高了情绪识别率。
To recognize emotion by physiological signals,facial and body expression,emotion self-report.60 subjects received emotion stimulus from fear,joy to relax.Signals of each emotion from 55 effective subjects were intercepted by 1 minute according to signal marks and GSR differentials.Parameters from these signals were firstly conducted emotional one-way anova and then stepwise discriminant to extract features to recognize emotions.Result showed that HRH、RV、R and T wave were more sensitive to emotions;7 features such as HFP,HR max,PNN50,LF/ HF,Ratio,LFP,Mean NN were extracted to construct discriminant functions: fuction1,fuction2 and Z1,Z2,Z3;the right recognition ratios were 88.0%,92.0%,80.0% for relax,joy and fear respectively and the total right recognition ratio was 86.7%.Its’ an effective method to recognize emotion by physiological signals,facial and body expression,emotion self-report.
出处
《生物医学工程研究》
2006年第3期141-146,共6页
Journal Of Biomedical Engineering Research
基金
江苏省自然科学基金资助项目(04KJB310171)
江苏大学高级人才科研启动基金资助项目(05JDG029)
关键词
情绪识别
情感计算
生理信号
特征提取
HRV
Emotion recognition
Affective computing
Physiological signals
Feature extraction
HRV