Learning State Assessment in Online Education Based on Multiple Facial Features Detection

Area of Science:

• Computer Science
• Human-Computer Interaction
• Educational Technology

Background:

• Online training often lacks effective supervision, leading to challenges in assessing learner engagement.
• Monitoring learner state is crucial for improving the effectiveness of online educational platforms.
• Existing methods may not adequately capture the nuances of learner attention and fatigue.

Purpose of the Study:

• To develop and validate an effective online learning state assessment approach.
• To combine multiple physiological and behavioral indicators for comprehensive learner state evaluation.
• To enhance the real-time monitoring capabilities for online learning environments.

Main Methods:

• Utilized blink detection via eye aspect ratio and blink frequency to assess eye fatigue.
• Implemented yawn detection using mouth aspect ratio with inner lip features for accurate mouth state analysis.
• Employed head pose estimation through 3D face model matching and Euler angle calculation to track head orientation.
• Processed all detections using 2D grayscale images for computational efficiency and real-time performance.

Main Results:

• The combined approach effectively differentiates various online learner states.
• Numerical analysis of blinking, yawning, and head pose changes accurately reflects learning engagement.
• Experimental results demonstrate the system's capability to evaluate online learning states.

Conclusions:

• The proposed method offers a robust and efficient solution for assessing online learning states.
• This technology can provide valuable support for the advancement and personalization of online education.
• Integrating behavioral cues like blinking, yawning, and head pose offers a promising direction for intelligent tutoring systems.