Measuring Student Attentiveness Using Eye-tracking and Visual-spatial Data Analytics

With the increasing trend toward e-learning, accurately assessing student attentiveness has become a critical need for effective education. In traditional classrooms, teachers rely on nonverbal cues to gauge student engagement, but online environments limit this ability, hindering real-time feedback that can help adapt instructional methods. To address this challenge, we proposed a student eye movements-based model to assess the student's attentiveness more effectively. For the student's attentiveness, the data is directly collected using ETVision glasses, and the features of eye movements are selected and preprocessed. To detect students' attentiveness and drowsiness, we propose a modified version of the Closed Eye Aspect Ratio (CEAR) model. Although the traditional model requires six eye landmark positions, in this paper, the proposed model requires only four eye landmark positions for the vertical dimension's eye landmarks. However, as virtual vision-mapped-to-real vision relation can be exposed in the Pixel Per Inch (PPI) measurement, the essential eye aspect ratio can be obtained from the vertical to the horizontal dimensions. The proposed approach is deployed in our experiments to analyze the students' eye-movement tracking and distraction behaviors. The analysis of the results has testified that the proposed model effectively detects students' attentiveness and drowsiness.