Real-Time Detection of Student Attention in Online Learning: A Scalable Solution for Crisis-Driven Remote Education

The COVID-19 pandemic has sharply highlighted the issues associated with keeping students engaged in online learning environments. In response, we introduce a real-time system for detecting and analyzing student attention during virtual classes. A real-time monitoring of students engagement should substitute the visual communication among students and educators in Face-to-Face (F2F) learning and subsequently allow the later to adjust their teaching behaviours during the Online classes. By mining the Students Attention Detection Dataset and Leveraging random forest along with feature engineering, Our model captures student behaviours like face orientation, hand gestures, phone usage, etc. in order to differentiate between attentive and inattentive students. Our study, demonstrates that features This research has implications beyond COVID-19, as online learning becomes more of a solution during natural disasters such floods, storms or other crises that impact traditional education. Our method provides a solution scalable to avoid learning disturbance in emergency conditions of remote learning. Early results show that real-time monitoring has significantly increased student engagement and are setting the foundation for more adaptable insights towards teaching during future disruptions.