基于CNN深度学习模型的大学生课堂行为检测研究  

Research on college students’classroom behavior detection based on CNN deep learning model

作　　者：左国才[1] 苏秀芝[1] 王海东[2] 吴小平[1] ZUO Guocai;SU Xiuzhi;WANG Haidong;WU Xiaoping(Hunan Vocational Institute of Software,Xiangtan Hunan 411100,China;Hunan University,Changsha 410082,China)

机构地区：[1]湖南软件职业学院,湖南湘潭411100 [2]湖南大学,长沙410082

出　　处：《智能计算机与应用》2020年第2期158-160,共3页Intelligent Computer and Applications

基　　金：湖南省教育科学规划课题研究成果(XJK19CZY018)。

摘　　要：为了切实提高职业院校教师课堂教学质量,对学生在专业课程课堂中表现出的各类上课行为进行检测分析,从而量化判断学生是否专注于课堂和教师的实际教学效果,达到提高教师课堂教学质量和督促学生专注课堂学习的目的。本文提出将卷积神经网络CNN深度学习模型应用于学生课堂行为检测识别,实现对学生是否专注课堂学习的行为进行分类。实验证明,该卷积神经网络能够对检测目标特征进行深度特征提取,并且对学生上课中的课堂行为检测取得良好的识别效果。In order to improve the classroom teaching quality of teachers in vocational colleges,the various classroom behaviors of students in professional courses are tested and analyzed,which could quantify whether students focus on the classroom and the actual teaching effect of teachers,therefore achieve the purpose of improving the classroom teaching quality of teachers and supervising students to concentrate on classroom learning.The paper proposes applying the Convolutional Neural Network in-depth learning model to the detection and recognition of students’classroom behavior,so as to classify whether the students are focused on classroom learning.Experiments show that the Convolution Neural Network can extract the depth features of the detected target features,and achieve good recognition results for students’classroom behavior detection in class.

关 键 词：卷积神经网络 深度学习 课堂行为 

分 类 号：TP311.52[自动化与计算机技术—计算机软件与理论]