摘要
计算机视觉技术发展为智能化教学提供了有力支撑。针对课堂行为自动识别问题,在分析学生课堂行为、识别流程和应用现状基础上,重点从目标检测、特征提取、行为分类三个方面综述了课堂行为识别方法,对比分析各方法优缺点,探讨了课堂行为识别在标准数据集构建、多目标检测、行为识别准确性等方面存在的问题与发展趋势。
The development of computer vision technology has provided favourable support for intelligent teaching. Aiming at the problem of automatic classroom behaviour recognition, on the basis of analysis of students’ classroom behaviour, recognition process and application status, classroom behaviour recognition methods are reviewed with emphasis on three aspects: target detection, feature extraction and behaviour classification, comparing and analysing the advantages and disadvantages of each method, and discussing the problems and development trends of classroom behaviour recognition in terms of standard data set construction, multi-target detection and behaviour recognition accuracy.
作者
贾轶钧
杨辉跃
JIA Yijun;YANG Huiyue(Army Logistics University,Chongqing 401331,China)
出处
《自动化与仪器仪表》
2022年第9期1-6,共6页
Automation & Instrumentation
基金
重庆市教委科技研究项目(KJQN202012903)。
关键词
课堂行为识别
目标检测
特征提取
行为分类
classroom behavior identification
target detection
feature extraction
behavior classification
