Algorithm for Student Behavior Detection Based on YOLOv8

Recently, major advancements have been achieved in the object detection algorithms for behavior detection, but detection for student behavior in low-resolution scenarios remains a great challenge. Aiming at the task of head-up rate detection in low-resolution scenarios, we added small object detection head, SPD-Conv module and NWD loss function to YOLOv8. Experiment has verified that our model has good detection performance for both large targets and low-resolution targets. This method has a mAP<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">50</inf> of 85.08% on the self-built dataset, exceeding the original YOLOv8n network by 2.77 percentage points.