Research on Hidden Mind-Wandering Detection Algorithm for Online Classroom Based on Temporal Analysis of Eye Gaze Direction

To enhance students’ self-management skills and improve the effectiveness of online learning, and to help teachers understand each student’s learning behavior, this paper proposes and designs an attention detection algorithm based on the temporal analysis of gaze direction. The algorithm aims to detect hidden distraction phenomena during online classes. The algorithm utilizes the FaceSSD face detection model and the Dlib-based facial landmark detection model to achieve the detection of students’ gaze direction. It extracts real-time time-series data regarding students’ gaze direction during class and applies a compressed and improved dynamic time warping algorithm for distraction detection. Furthermore, to enhance the real-time feedback capability of the system, a top student strategy is proposed to reduce the computational workload. Experimental results demonstrate that the proposed algorithm can quickly and effectively detect students’ distraction phenomena, including daydreaming, distraction, and hidden activities like using mobile phones in blind spots of the camera’s visual capture. This research is significant for helping teachers evaluate students’ performance in online classrooms and enable intelligent monitoring of teaching quality.