EmoVision: Real-Time Engagement Detection in MOOCs

The high dropout rates, limited pedagogical staff, and isolation experienced by learners are significant challenges in Massive Open Online Courses (MOOCs). Traditional methods for assessing learner engagement often fall short in providing immediate and objective insights. This paper introduces a real-time engagement and emotion detection system that leverages computer vision and machine learning (ML) technologies. The system employs Convolutional Neural Networks (CNNs) for facial emotion recognition and utilizes Single Shot Multibox Detector (SSD) and Haar Cascade classifiers for face detection. By identifying six key emotions and calculating an engagement index, the system provides instant feedback to educators, MOOC designers, profiling systems, and recommender systems. This enables personalized and adaptive learning experiences, addressing the impersonal nature of MOOCs and enhancing overall learner engagement and outcomes. This technology represents a significant advancement in online education, aiming to reduce dropout rates and improve learner satisfaction.