微表情类别与区域间关联度的分析方法研究

doi:10.3778/j.issn.1002-8331.1907-0298

计算机工程与应用 ›› 2020, Vol. 56 ›› Issue (19): 146-151.DOI: 10.3778/j.issn.1002-8331.1907-0298

微表情类别与区域间关联度的分析方法研究

张延良，卢冰，蒋涵笑，洪晓鹏，赵国英，张伟涛

1.河南理工大学物理与电子信息学院，河南焦作 454000
2.西安交通大学电子与信息工程学院，西安 710049
3.奥卢大学机器视觉和信号分析研究中心，芬兰奥卢 FI-90014
4.西安电子科技大学电子工程学院，西安 710071

出版日期:2020-10-01 发布日期:2020-09-29

Study on Investigating Correlation Between Categories of Micro-expression and Corresponded Regions

ZHANG Yanliang, LU Bing, JIANG Hanxiao, HONG Xiaopeng, ZHAO Guoying, ZHANG Weitao

1.School of Physics and Electronic Information, Henan Polytechnic University, Jiaozuo, Henan 454000, China
2.School of Electronic and Information Engineering, Xi’an Jiao Tong University, Xi’an 710049, China
3.Center for Machine Vision and Signal Analysis, University of Oulu, Oulu FI-90014, Finland
4.School of Electronic Engineering, Xidian University, Xi’an 710071, China

Online:2020-10-01 Published:2020-09-29

摘要/Abstract

摘要：

为了揭示微表情类别与面部各区域之间的关联度，提出一种局部区域划分和分析方法。该方法首先根据各类微表情对应动作单元所在的位置确定出人脸中最为关键的7个局部区域，然后采用面部形变鉴别模型自动检测出面部49个关键点坐标，再根据这些点确定上述局部区域的4个边界。最后，分别提取各区域的时空模式特征并进行微表情分类，通过各区域对各类微表情的分类准确率来分析两者之间的关联度。实验结果表明，局部区域划分方案科学合理，微表情“惊奇”及“厌恶”与眼睛区域、微表情“高兴”与嘴巴区域、微表情“压抑”与下巴区域有较高的关联度。

关键词: 微表情识别, 局部区域, 动作单元, 面部形变鉴别模型, 分类准确率

Abstract:

In order to reveal the correlation between the micro-expression categories and the local regions, the method for partitioning and analyzing local region is proposed. Firstly, based on the location of action units related to various micro-expressions, seven crucial local regions of face are determined. Then, 49 facial point coordinates are detected automatically through facial deformation identification model and the four boundaries of the aforementioned local regions are established according to these coordinates. The spatio-temporal pattern features of each region are extracted and the micro-expressions are classified. Finally, the correlation between the micro-expression categories and the local regions are obtained by the recognition accuracy of each region for every micro-expression category. The experimental results show that the proposed scheme for partitioning local region is precise and reasonable; there is a highly correlation between the eye area and surprise or disgust. The similar correlations also lie in the mouth area and the happiness, the chin area and the depression.

Key words: micro-expression recognition, local region, action units, facial deformation identification model, recognition accuracy

张延良，卢冰，蒋涵笑，洪晓鹏，赵国英，张伟涛. 微表情类别与区域间关联度的分析方法研究[J]. 计算机工程与应用, 2020, 56(19): 146-151.

ZHANG Yanliang, LU Bing, JIANG Hanxiao, HONG Xiaopeng, ZHAO Guoying, ZHANG Weitao. Study on Investigating Correlation Between Categories of Micro-expression and Corresponded Regions[J]. Computer Engineering and Applications, 2020, 56(19): 146-151.

[1]	温杰彬，杨文忠，马国祥，张志豪，李海磊. 基于Apex帧光流和卷积自编码器的微表情识别[J]. 计算机工程与应用, 2021, 57(4): 127-133.
[2]	张力为，王甦菁，段先华. 细粒度分层时空特征描述符的微表情识别方法[J]. 计算机工程与应用, 2021, 57(13): 154-160.
[3]	张人，何宁. 微表情识别研究综述[J]. 计算机工程与应用, 2021, 57(1): 38-47.
[4]	刘振，王甦菁，李擎. 基于多任务中级特征个性化学习的微表情识别[J]. 计算机工程与应用, 2019, 55(18): 151-154.
[5]	郑伟，陈彦江. 基于局部区域信息的水平集医学图像分割方法[J]. 计算机工程与应用, 2010, 46(31): 209-211.
[6]	常军,吴锡生. 基于图像子块加权缩小的自适应修正算法[J]. 计算机工程与应用, 2009, 45(7): 181-183.
[7]	李楠^1，2，肖克炎²，李源³，陈析璆¹，邹伟². 一种三角形折叠网格模型简化的改进算法[J]. 计算机工程与应用, 2009, 45(34): 192-194.

微表情类别与区域间关联度的分析方法研究

Study on Investigating Correlation Between Categories of Micro-expression and Corresponded Regions

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 7

编辑推荐

Metrics