Research on the Detection and Analysis of Students’ Classroom Behavioral Features Based on Deep CNNs

Abstract

The continuous development of artificial intelligence technology has caused the traditional education model to begin to change as well. The new educational model of applying intelligent monitoring devices to modern classrooms, thus assisting teachers in teaching students, has been successfully favored by major universities. Based on this, this research attempts to improve the target detection algorithm in the field of deep learning, and proposes a method of detecting and analyzing students' classroom behavior based on deep convolutional neural network, and applies the method to detect students' classroom behavior characteristics, aiming to improve the method of detecting and analyzing students' classroom behavior, assist teachers in student management with intelligent monitoring devices, and thus improve the quality of teaching. The study firstly investigated the core steps of the target detection algorithm, and secondly improved the YOLOv5s algorithm by adding feature fusion layer and reimaging module, and finally obtained the Ghost-4D-YOLOv5s network model. The experimental results show that the Ghost-4D-YOLOv5s model performs well in terms of precision, recall, F1 value and average precision mean. When the number of samples to be tested is 350, the performance of the model achieves 58.2% precision, 59.6% recall, 62.5% F1 value and 62.8% average precision mean value. This indicates that the model constructed in this study has good performance in detecting the behavioral characteristics of students in the classroom. It can be concluded that the model constructed in this study can have a good performance in the detection of students' classroom behavioral characteristics. Through this model, teachers can better manage their students and improve the quality of teaching. In addition, this study highlights the potential of deep learning in the field of education and contributes to the development of intelligent educational models.