Neural network-based protective mask detection platform

This study suggests a real-time face mask identification system that uses the VGG16 architecture for mask classification and Multi-task Cascaded Convolutional Networks (MTCNN) for accurate face detection in order to address health and safety concerns during the COVID-19 pandemic. The system combines temperature screening and mask detection for all-encompassing health monitoring, enhanced by thermal imaging. Using precise face bounding boxes supplied by MTCNN, the VGG16 model—pre-trained on ImageNet—was optimized for binary classification using a dataset of 7,553 photos of masked and unmasked faces. The system exhibits resilience to changing angles, lighting conditions, and partial occlusions, achieving an accuracy rate of 98%. While image augmentation and zoomed face regions increase detection reliability in complex settings, including surveillance film from public spaces, thermal detection recognizes people with increased temperatures. This technology supports health activities in public locations like malls, airports, and workplaces by automating safety enforcement through the combination of temperature screening and mask compliance monitoring. Future developments will concentrate on improving deployment on edge devices and extending features to incorporate more safety and health precautions.