Analysis of Facial Expression to Estimate the Levels of Engagement in the Online Lectures

In online education, understanding students' engagement is essential for improving learning outcomes. This paper proposes a method to estimate students' attentional states through facial expression analysis. By recording reaction times (RT) to an irrelevant auditory stimulus during online lectures, we gauge students' focus levels. Facial features were extracted using OpenFace, and LightGBM models were trained to predict RT from these features. The findings indicate a significant relationship between facial expressions and attentional states, highlighting potential applications in adaptive online learning environments.