Published Paper Details:

This paper suggests an interactive dashboard that uses deep learning methodologies to enhance the online classroom environment through student emotion analysis. The main challenge in digital learning is occluded or partially hidden facial data, for which CNNs and Haar cascade come into play. Haar cascade is responsible for facial feature detection and extraction by which CNNs can classify emotions. Real-time emotion prediction is visualized on the interactive dashboard, combining inputs such as the subject, course, and instructor. Analytics on the dashboards are represented in charts and graphs and help inform teachers about student responses and how they should tweak their teaching to better engage students. It also assists in identifying students who may require extra support. The system was validated on the CK+ dataset and was found to be highly accurate in classifying emotions such as happiness, sadness, and surprise. Other than the occasional instances in which students' faces were not visible, the system proved to be robust and consistent. With real-time emotion-based feedback, this dashboard has transformed the dynamics of a virtual classroom into one that fosters adaptive and engaging learning, promoting the optimum achievement of educational objectives.