摘要
提高教育质量是培养创新型人才的关键。而传统课堂中,教育者掌握学生行为要通过课堂观察和随机提问等人工的方式进行,此类方式存在课堂信息传递与反馈滞后等问题。为此,基于YOLOv8深度学习算法,构建了学生课堂行为检测模型。通过在真实课堂环境中采集数据,并对其进行标注、格式转换等处理,将处理好的数据用于训练以此构建检测模型。该模型对写字、端坐、玩手机、睡觉、站立、低头6种学生课堂行为进行识别。研究表明,检测结果精确度达到83.3%,验证了此模型的有效性,实现了对学生课堂行为自动化识别和分类,以此检测结果辅助教育者判断学生学习情况并做出教学决策。
Improving the quality of education is the key to cultivate innovative talents.In the traditional classroom,educators master students'behavior through manual methods such as classroom observation and random questioning.Such methods have problems such as lagging classroom information transmission and feedback.Therefore,based on the YOLOv8 deep learning algorithm,a student classroom behavior detection model is constructed.By collecting data in the real classroom environment,labeling and format conversion,the processed data is used for training to construct the detection model.The model identifies six kinds of students'classroom behaviors,including writing,sitting,playing with mobile phones,sleeping,standing and lowering head.The research shows that the accuracy of the test results reaches 83.3%,which verifies that this model realizes the automatic identification and classification of students'classroom behaviors,and assists educators to judge students'learning situation and make teaching decisions.
作者
王璇
张桂杰
WANG Xuan;ZHANG Gui-jie(Jilin Normal University,Siping 136000,China)
出处
《电脑与电信》
2024年第8期26-32,共7页
Computer & Telecommunication
基金
吉林省科技厅科学技术项目“多维时间序列数据相似性度量方法研究”,项目编号:20230101243JC。
关键词
目标检测
YOLOv8
学生课堂行为检测
模型构建
target detection
YOLOv8
students'classroom behavior detection
model construction
