Facial emotion-based student's attention in online learning using deep learning

Despite the growing popularity of online courses, it is still difficult to guarantee students' participation. Online students typically struggle to stay motivated and concentrated, in contrast to students in regular classrooms who benefit from direct supervision and motivation. Facial expression recognition is built into the framework so that the emotional state of the learner may be monitored and evaluated in real-time. To detect and evaluate student interest in the present moment, a deep learning-based strategy is proposed, incorporating deep learning models like VGG-19 and ResNet-50 for face expression identification. An attention index (AI) is determined through the analysis and classification of facial expressions, allowing for the forecasting of 'attentive' and 'inattentive' states. Accuracy is very high, with VGG-19 being the most accurate model at 93.11%. These results show the potential of deep learning as a tool for measuring and retaining students' interest.