The rapid expansion of online education, especially following the COVID-19 pandemic, has transformed e-learning platforms into a dominant mode of academic delivery. Popular platforms like Coursera, edX, Udemy, and NPTEL serve millions of learners globally. However, these systems lack the ability to understand and respond to students’ emotional states, resulting in a passive learning experience and contributing to high dropout rates. Unlike traditional classrooms, where instructors can interpret student expressions and adapt teaching methods accordingly, online platforms fail to provide such real-time feedback mechanisms. This project introduces EmotionLearn, an Affective Learning System inspired by the principles of Affective Computing proposed by Rosalind Picard. The system integrates real-time facial emotion recognition into a web-based learning environment to enhance student engagement and learning effectiveness. By leveraging advancements in deep learning and computer vision, particularly Convolutional Neural Networks (CNNs), the system detects key emotional states such as Happy, Neutral, Confused, Sad, and Angry during video-based learning sessions.
Pathan et al. (Fri,) studied this question.