Smart E-learning: Enhancement of Human-Computer Interactions Using Head Posture Images

This paper proposes a novel e-learning system that incorporates human-computer interaction data to build a smart e-learning system. A supervised image segmentation algorithm is used to detect the face and hair of students in head posture images. A simple and effective human presence detection and gaze direction estimation method is then developed based on changes in the face and hair information. First, the proposed algorithm is tested using 10 different students with seven different head postures each and 92% of the head postures are identified accurately. Second, the method is applied to real time video sequences containing 80 frames that lasted 400 seconds, which are acquired using an integrated web camera, and similar results are obtained. Finally, human-computer interaction data, which is an indicator of student attention, is calculated based on the human presence and gaze direction over time. The experimental results show that the proposed approach enhances human-computer interactions for e-learning systems and helps us to evaluate student performance.