计算机应用与软件2017，Vol.34Issue(1)：165-169,5.DOI:10.3969/j.issn.1000-386x.2017.01.030
基于差分定位与光流特征提取的微表情识别
MICRO EXPRESSION RECOGNITION BASED ON DIFFERENTIAL POSITIONING AND OPTICAL FLOW FEATURE EXTRACTION
摘要
Abstract
Existing micro expression recognition algorithm has the problems of redundancy in extracting feature and long processing time.In view of this, we proposed a micro expression recognition method which is based on differential positioning and optical flow feature extraction.First, it applies image sequence difference on sensitive parts of facial expression, delimits motion area on face by differential projection value.Then it calculates principal component feature of optical flow analysis data in motion area.Finally, it uses support vector machine to determine the number of facial movement unit and recognises the category of micro expression.Experimental results showed that compared with conventional recognition technology, this algorithm has significant improvement on efficiency and accuracy of micro expression recognition.关键词
微表情识别/光流分析/图像差分/主成分分析/面部行为编码系统Key words
Micro expression recognition/Optical flow analysis/Image difference/Principle component analysis/Facial action coding system分类
信息技术与安全科学引用本文复制引用
许刚,赵中原,谈元鹏..基于差分定位与光流特征提取的微表情识别[J].计算机应用与软件,2017,34(1):165-169,5.基金项目
中央高校基本科研业务费专项项目(2015XS17). （2015XS17）
