摘要
针对同步课堂教学中存在的远程学情感知问题,本文对课堂投入度计算进行了探索。结合Faster R-cnn深度学习技术,文章提出了一种依据学生头部姿势、手部姿势、身体姿势实现行为识别和课堂投入度计算的方法,结果显示具有较好的目标检测率、识别准确率和目标跟踪性能;同时,采用十个课堂教学视频进行投入度自动化评价的实验与分析,结果证明本文提出的课堂投入度自动分析技术可以通过计算课堂中个人和班级的实时投入度和阶段性投入度,为教师提供良好的远程学情感知方式。
Aiming at the problem of emotional knowledge in distance learning in synchronous classroom teaching,this paper explores the calculation of classroom engagement.Combined with Faster r-cnn deep learning technology,this paper proposes a method to realize behavior recognition and classroom engagement calculation according to students’head posture,hand posture and body posture.The results show that it has good target detection rate,recognition accuracy and target tracking performance;At the same time,ten classroom teaching videos are used for the experiment and analysis of automatic evaluation of classroom engagement.The results prove that the automatic analysis technology of classroom engagement proposed in this paper can provide teachers with a good way of distance learning perception by calculating the real-time and phased engagement of individuals and classes in the classroom.
作者
邓伟
夏磊杰
刘清堂
张思
魏艳涛
Deng Wei;Xia Leijie;Liu Qingtang;Zhang Si;Wei Yantao
出处
《高等继续教育学报》
2022年第6期15-24,共10页
Journal of Continuing Higher Education
基金
教育部人文社会科学研究规划基金项目“基于远程课堂学习情绪计算的同步课堂教学干预机制研究”(20YJA880009)
华中师范大学国家教师发展协同创新实验基地建设研究项目(CCNUTEIII 2021-09)
高等师范院校基础教育工作研究会“‘互联网+’教育教学改革实践研究——以远程学情职能感知促进教师在线教学能力提升”。
关键词
学习投入度
图像识别
同步课堂
远程学情感知
投入度分析
learning engagement
image recognition
synchronous classroom
distance learning emotion perception
input analysis
