Student Classroom Behavior

Research on student classroom behavior focuses on automatically detecting and classifying student actions (e.g., hand-raising, reading, using phones) from video data to improve teaching effectiveness and analyze student engagement. Current studies utilize deep learning models, particularly object detection architectures like YOLOv7 and spatio-temporal networks like SlowFast, often incorporating attention mechanisms to enhance accuracy and address class imbalances. The development of publicly available datasets is a key focus, enabling researchers to benchmark and improve algorithms for more accurate and robust behavior detection. This work has the potential to provide valuable insights into student learning and inform pedagogical strategies.