Spartan Face Mask Detection and Facial Recognition System

Area of Science:

• Computer Science
• Artificial Intelligence
• Biomedical Engineering

Background:

• Face masks are crucial for preventing airborne disease transmission, as recommended by the WHO.
• COVID-19 mandates have led to widespread mask usage, complicating facial recognition systems.
• Mask effectiveness depends on type and proper positioning, adding complexity to identification.

Purpose of the Study:

• To develop a robust face detection and recognition system capable of handling masked individuals.
• To address challenges in identifying people in public spaces during health crises.
• To create a cost-efficient solution for enterprises and educational institutions.

Main Methods:

• Utilized stacking ensemble deep learning algorithms for comprehensive facial analysis.
• Employed Convolutional Neural Networks (CNN), AlexNet, and VGG16 for feature classification.
• Integrated a FaceNet-based facial recognition pipeline for identity verification.

Main Results:

• The Spartan system effectively performs mask detection, type classification, and position classification.
• Achieved accurate identity recognition even with individuals wearing face masks.
• Demonstrated the system's capability to function on edge devices.

Conclusions:

• The proposed system offers a viable solution for secure identification in mask-mandated environments.
• The technology enhances security and operational efficiency for organizations.
• The system is adaptable for deployment on edge devices, providing flexibility and cost-effectiveness.