摘要
新冠疫情导致全球在线教育异军突起,如何打破时间、空间的限制实时监督学生的在线学习状态随时调整教学策略进而提高学习效率具有重要意义。基于2D与3D图像处理技术,提出了一种在线评估学习状态的方法,将学习者的学习状态分为无人、多人、用户未授权、分心以及疲劳五种。利用AdaBoost算法与ResNet模型实现人脸检测和识别,并加入质心跟踪算法提高人脸识别检测效率;利用RGB-D图像实时获取人脸三维模型,通过EPNP算法获取学习者头部姿态进而评估学习姿态;提取学习者眼睛和嘴巴的实时图像特征,获取学习者眼睛与嘴巴纵横比值,实现学习者学习疲劳状态的实时预警。测试表明学习者的身份识别、姿态评估以及疲劳预警有效可行,为提高在线学习的质量提供了积极思路。
In light of the emerging global online education resulted from the outbreak of COVID-19,how to break the time and space limits to supervise students’online learning conditions on timely units and adjust teaching strategies freely to facilitate their learning efficiency is of paramount significance.Based on the processing technologies in 2D and 3D images,we put forward an online evaluation method of learning conditions,which divides the learning condition of learners into five categories:No Person,More than One Person,Failed Authorization from the User,Distraction and Fatigue.AdaBoost algorithm and ResNet model are used to realize face detection and recognition,and centroid tracking algorithm is added to improve the efficiency of face recognition detection.RGB-D image can be adopted to acquire 3D face models at any time while EPNP algorithm can help collect their head gesture and evaluate their learning conditions.Real-time image characteristics about their eyes and mouth is extracted to get an aspect ratio between the two parts,making it possible to achieve infinite warning of their fatigue in learning.As suggested in the test,the practice proves feasible in recognizing the learner’s identity,evaluating his gestures and warning of their fatigue,which provides a positive thought on improving the quality of online learning.
作者
姚远
徐晶晶
朱小倩
YAO Yuan;XU Jing-jing;ZHU Xiao-qian(School of Physical Science and Technology,Central China Normal University,Wuhan 430079,China)
出处
《计算机技术与发展》
2021年第12期128-134,共7页
Computer Technology and Development
基金
2019年湖北省知识产权运用示范工程(高校院所)后补助项目(35)。
关键词
图像处理
在线学习
身份识别
姿态评估
疲劳预警
image processing
online learning
identity recognition
posture assessment
fatigue warning
