摘要
传统专家随堂听课方式的教学质量评价在人员、时间方面花费很大。对此,文章基于课堂视频,采用人工智能的方法,对学生状态进行分析并对指标进行量化:通过深度学习算法对学生数量进行检测,通过机器学习算法对学生位置分布进行分析、对学生人脸关键点进行检测并对学生表情进行分类。评价内容主要包含学生数量检测及位置分布、学生表情及姿态识别,以及对学生个体、整体的统计分析等。该课堂评价体系具有信息反馈的实时性和高效性,可辅助教师改进授课方式。
The teaching quality evaluation by the traditional way of experts visiting is costly in terms of staff and time.In this regard,the paper analyzed students’statuses and quantified the indicators based on class videos using the artificial intelligence method:the number of students was detected by the algorithm of deep learning;the location distributions of students was analyzed by the algorithm of machine learning,key points of students’faces were detected and students’face expressions were classified.The evaluation content mainly included three parts of the detection of students’number and the distributions,the recognition of students’face expressions and gestures,and the statistical analyses of the individual and the whole.The established classroom evaluation system had the real-time and high-efficiency of information feedback.
作者
贾鹂宇
张朝晖
赵小燕
闫晓炜
JIA Li-yu;ZHANG Zhao-hui;ZHAO Xiao-yan;YAN Xiao-wei(School of Automation,University of Science and Technology Beijing,Beijing,China 100083)
出处
《现代教育技术》
CSSCI
北大核心
2019年第12期82-88,共7页
Modern Educational Technology
基金
北京市人才培养共建项目“基于信息技术的课堂教学质量控制研究”（项目编号：GJ1804）的阶段性研究成果
关键词
人工智能
课堂评价
表情识别
artificial intelligence
classroom evaluation
expression recognition
