摘要
应用Coiflet 1小波对脉搏波信号进行小波变换,利用重构的小波第三层高频信号的极大值来确定原始信号主波峰值的范围,进而定位原始信号主波峰值的位置并以此位置为参照点确定潮波、重搏波峰值的位置,在此基础上提取脉搏信号的时域特征,对情感状态进行分类.实验结果表明,脉搏率均值等特征对悲伤、愤怒和恐惧3种情感的总体分类正确率均能达到65%以上.
Coiflet 1 is adopted to do wavelet transform,the maximum of the third layer high-frequency signal in reconstructed wavelet is used to determine the major wave crest value range of the original signal,and then locate the position of the major crest value of the original signal.Referring to the location of the major wave crest value,locations of the tidal wave and the dicrotic wave crest values can be determined,on this basis,time-domain features of pulse signal are extracted and the emotion status are categorized.The stimulation results show that recognizing three affective status（sadness,anger and fear） from features such as pulse rate average value,the recognition rate can reach more than 65%.
出处
《西南师范大学学报（自然科学版）》
CAS
CSCD
北大核心
2010年第3期243-246,共4页
Journal of Southwest China Normal University(Natural Science Edition)
基金
国家自然科学基金(60873143)
国家重点学科基础心理学科研基金(NKSF07003)
关键词
脉搏波
离散小波变换
脉搏率
情感识别
pulse
discrete wavelet transform
pulse rate
emotion recognition
