基于人脸检测的课堂出勤率统计方法研究  

Research on Classroom Attendance Statistics Method Based on Face Detection

作　　者：杨长凡 吕梦鸽 周泽宇 张子腾 赵彤阳 赵桐 Yang Changfan;Lu Mengge;Zhou Zeyu;Zhang Ziteng;Zhao Tongyang;Zhao Tong(China University of Mining and Technology,Beijing 100083)

机构地区：[1]中国矿业大学,北京100083

出　　处：《南方农机》2021年第11期172-175,共4页

基　　金：中国矿业大学(北京)大学生创新训练项目“基于人脸检测的课堂出勤率统计方法研究”(C202004927)。

摘　　要：为了寻找一种高效的课堂考勤方法,本文借助课堂监控视频,提出了一种基于人脸识别算法的多等级评价课堂考勤方法。首先,针对两门不同性质的课程监控视频,提取若干张视频图像;然后,利用Retinaface算法对图片中的人脸进行检测;最后,基于图片中人脸检测的结果,提出课堂考勤的四个评价等级,可将学生的课堂出勤、听课情况等信息综合反馈给授课教师。经计算得出,电机课和电工课的最终得分分别为4.78分和4.55分,均为优秀,并从分数的角度体现了学生对专业核心必修课的重视程度。In order to find an efficient classroom attendance method,this paper uses classroom surveillance video to propose a multilevel evaluation classroom attendance method based on face recognition algorithm.First,for two different types of course surveillance videos,extract several video images.Then,the Retinaface algorithm is used to detect the face in the picture.Finally,based on the results of face detection in the pictures,four evaluation levels of classroom attendance are proposed,which can comprehensively feed back information such as students’classroom attendance and class attendance to the instructor.After calculation,the final scores of electrical engineering and electrical engineering are 4.78 and 4.55 respectively,both of which are excellent,and from the point of view of scores reflect the degree of importance students attach to the core compulsory courses of the major.

关 键 词：人工智能 人脸检测 Retinaface算法 出勤率统计 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]