摘要
在基于心电的情感识别中,体现不同情感状态的特征是提高情感识别率的基础。采用最优小波去噪心电信号,进行P波、QRS波、T波检测和能量计算,然后提取特征进行情感分类,并分析了P-QRS-T波能量在不同情感状态下的变化趋势和对情感状态的敏感性。实验结果表明,P-QRS-T波能量变化能体现情感状态的变化,对高兴敏感,识别率可达96%;同时最优小波去噪能有效地提高情感状态识别率。
In the emotion recognition based on ECG,the features which express different emotion status well are the basis of improving recognition ratio.In this paper,after ECG denoised by optimal wavelet,detect P,QRS and T waves and compute their energy value,then extract feature to classify emotions,also analyze the trend of variance in P-QRS-T waves＇ energy and its sen- sitivity to emotion status under different emotions.The main result shows that P-QRS-T waves＇ energy changes with the variance of emotion status and is sensitive to Joy with recognition ratio of 96%.At the same time,the emotional recognition ratio increases with the optimal wavelet denoising.
出处
《计算机工程与应用》
CSCD
北大核心
2009年第8期213-215,共3页
Computer Engineering and Applications
基金
国家自然科学基金No.60873143~~
关键词
情感识别
特征提取
最优小波去噪
P—QRS—T波检测
emotion recognition
feature extraction
optimal wavelet denoising
P-QRS-T waves detection
