基于面部多特征融合的学生网课疲劳检测研究  

Research on Fatigue Detection of Students Online Course Based on Facial Multi-Feature Fusion

作　　者：黄丽凤 陈建华 HUANG Li-feng;CHEN Jian-hua(School of Software,Quanzhou University of Information Engineering,Quanzhou Fujian 362000,China)

机构地区：[1]泉州信息工程学院软件学院,福建泉州362000

出　　处：《计算机仿真》2023年第9期242-246,267,共6页Computer Simulation

基　　金：2021年度福建省中青年教师教育科研项目(JAT210551)。

摘　　要：随着远程教育的不断发展,网课成为了一种趋势。为了进一步监管学生网课学习时的疲劳状态,保障网课学习质量,研究了学生网课疲劳检测算法。利用OpenCV结合PHOG算法提取学生眼部和嘴巴特征点,采用眼部自适应闭眼阈值和基于欧拉角的特征校正算法实时计算学生眼部和嘴部的纵横比Er值和Mr值,得到反映学生疲劳状态的眼睑闭合时间占比值、眨眼频率、打哈欠频率,并为这三个特征指标设置权重,通过计算加权和融合,得到学生的疲劳值。据此判断学生疲劳等级并提出警示信息。实验表明,上述方法能够实时地反映学生的疲劳状态,学生疲劳判定的正确率达到95.5%以上,具有良好的实时性和辨别率。With the continuous development of distance education,online courses have become a trend.In order to further supervise the fatigue state of students during online course learning and ensure the quality of online course learning,the fatigue detection of students online course based on facial multi-feature fusion was studied.In this paper,OpenCV combined with PHOG algorithm was used to extract the key feature points of students'faces,and Er and Mr Values of the aspect ratio of students'eyes and mouth were calculated in real time by using the adaptive threshold of eye closure and the feature correction algorithm based on Euler Angle.PERCLOS values,blink frequency and yawn frequency reflecting students'fatigue state were obtained,and weights were set for these three indicators.By calculating the weight and fusion,the fatigue value of students was obtained.According to this,the fatigue levels of students were judged and the warning information was put forward.The experiment shows that the method can reflect the fatigue state of students in real time,and the correct rate of student fatigue determination is more than 95.5%,which has good real-time performance and discrimination rate.

关 键 词：网课 疲劳检测 多特征融合 眼睑闭合时间占比 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]