摘要
随着谎言检测的深入探究,微表情识别研究得到了很大的关注,心率变化作为体现情绪变化的另一种形式,也可能对微表情识别有一定的影响。研究目的是找寻心率变化和微表情的关系。利用欧拉视频放大算法,实现非接触式心率提取,并对目前已公开的微表情数据库CAS(ME)^2数据进行分析。通过测录视频进行实验分析,提出的心率提取算法可得到95%的准确率,对数据库数据进行心率检测的结果与表情数据进行对比,可以得到恐惧、愤怒、悲伤等情绪与心率变化存在一定关系。
With the deep exploration of lies detection,micro-expression recognition research has been a great concern.The changes of heart rate as a manifestation of emotional changes in another form may also have a certain impact on micro-expression recognition.The purpose of this paper is to find the relationship between heartbeat and micro-expression.We use Euler video amplification algorithm to achieve non-contact heart rate extraction,and analyze the data in micro-expression database CAS(ME)^ 2 which has been published currently.By recording videos for experimental analysis,the result of non-contact heart rate extraction can obtain 95% accuracy.And comparing the heart rate results and the expression data in the database,we can get that there is a certain relationship between the emotions which are fear,anger and sadness with the changes of heart rates.
作者
张树诚
王笑梅
ZHANG Shu-cheng;WANG Xiao-mei(Shanghai Normal University,Shanghai 200030,China)
出处
《计算机仿真》
北大核心
2019年第3期453-457,共5页
Computer Simulation
