Research On Intelligent Recognition Technology Of College Students’ Classroom Behavior Based On Improved SSD

The digital campus is a digital and informationized campus environment,and its purpose is to improve the quality of education and teaching and realize the modernization of education and teaching models.Among them,teaching video information is one of the important educational resources in the digital campus,and how to effectively use these resources is one of the problems to be solved in the construction of the digital campus.Aiming at this problem,this paper proposes an improved SSD target detection algorithm to intelligently analyze the state of students in the video.First of all,this paper selects one-semester classroom teaching videos of a university as the research object,and uses more than 10,000 1080 p video frames as the training data set.Aiming at the problem that the deep learning SSD model has poor detection results for students’ four behavioral states of listening to lectures,sleeping,playing with mobile phones,and taking notes,a Res Net residual network and a dual attention mechanism are introduced into the VGG module to improve the feature extraction ability of student states in video frames;A FPN detection model was built to improve the accuracy of image recognition for students’ note-taking categories and low recognition accuracy of small targets in the back row.And based on this model,the classroom student behavior recognition system is implemented.At the same time,in order to cope with the traffic fluctuations caused by the different number of classes in different periods of the college classroom,the micro-service architecture,docker and k8 s are used to achieve high concurrency and elastic expansion of the system.The experimental results show that the improved SSD model has a greater improvement in feature extraction ability and small target recognition accuracy than the original SSD,and the m AP has increased from the original 51.62%to 63.75%.The research results of this project can provide real-time information reference for the teaching management of colleges and universities,and have strong practicability.