An integrated facial recognition system for classroom resource optimization using MobileNet and SSA-SVM

Area of Science:

• Artificial Intelligence
• Educational Technology
• Computer Vision

Background:

• Efficient utilization of educational resources is crucial for academic institutions.
• Real-time monitoring and attendance tracking are key to optimizing classroom management.
• Existing systems may lack the accuracy and speed required for dynamic educational environments.

Purpose of the Study:

• To develop and evaluate an integrated facial recognition system for enhanced university classroom management.
• To improve resource allocation and utilization through real-time monitoring and attendance tracking.
• To assess the system's performance in terms of recognition accuracy and real-time processing capabilities.

Main Methods:

• Utilizing MobileNet for feature extraction combined with a SSA-SVM classification model.
• Implementing a depthwise separable convolutional network with an inverted residual module for precise feature extraction.
• Testing the system across institutions in Henan Province for performance evaluation.

Main Results:

• Achieved a 3.47% increase in face detection accuracy and a 7.05% increase in recognition rate compared to baseline methods.
• Reached an overall accuracy of 98.13% in face recognition.
• Demonstrated 93.61% accuracy at 125 frames per second in classroom settings, outperforming existing methods in efficiency.

Conclusions:

• The integrated facial recognition system effectively supports efficient classroom management and resource optimization.
• The proposed deep learning model shows significant improvements in accuracy and real-time processing for educational applications.
• This technology offers a viable solution for enhancing educational resource utilization in academic settings.