Face Mask Detection using CNN: A Fusion of Cryptography and Blockchain

The global COVID-19 pandemic has led to an urgent need for government intervention to prevent its spread. Scientific evidence has confirmed the effectiveness of mask-wearing in reducing virus spread. However, enforcing mask mandates in public spaces presents notable monitoring hurdles, particularly as facial recognition technology is impeded by face coverings. With many organizations relying on facial recognition for employee authentication, security and authentication become critical, especially for IoT systems. This article uses a Convolutional Neural Network (CNN) model to accurately identify mask-wearing individuals and introduces a secure user authentication mechanism between the node and the access point. This authentication mechanism consists of three phases. (i) Identification Phase at the Access Point Level: A novel hybrid biometric pattern, merging password and image features, is employed to strengthen user authentication security through a fusion approach. (ii) Secure Communication: Utilizing blockchain technology and AES cryptography ensures the secure transmission of these patterns between the node and the access point. (iii) Matching Phase at the Node Level: A newly proposed method verifies authenticity by comparing the combined image and password features with database records during the development phase. The experimental results demonstrate its outstanding performance, achieving 99% accuracy, 99% recall, 100% precision, and 98.9% F1 score. These results suggest that the proposed approach holds promise as an effective and secure solution for identifying individuals wearing masks while ensuring reliable authentication in various environments.