Student Classroom Behavior Detection Algorithm Based on Improved YOLOv8n

Currently, object detection technology is becoming more and more mature, but there are still a lot of challenges in recognizing students' classroom behaviors. In order to address the problems of inaccuracy and high computation of existing models in the process of student classroom behavior recognition, this paper adopts the improved Yolov8n model to detect and recognize student classroom behavior. Based on Yolov8n, the method adds an efficient shuffle attention to increase the ability of feature extraction and improve the model recognition accuracy; secondly, the bounding box loss function is optimized to improve the model's localization ability. The experimental results show that the mAP50 and mAP50-95 metrics of the proposed model on the student classroom behavior dataset are 99.9% and 95.4%, respectively. The proposed model can achieve the detection and identification of classroom behavior more quickly and accurately with lower computing cost, and can realize the dynamic and scientific identification of students' classroom learning.