A Parallel Convolutional Neural Network for Accurate Face Mask Detection in the Fight Against COVID-19

Abstract

The rapid spread of infectious diseases, such as COVID-19, has highlighted the critical need for reliable and efficient face mask detection systems. This study proposes a novel parallel hybrid convolutional neural network (CNN) architecture that integrates VGG16 and MobileNetV2 to enhance feature extraction and classification accuracy. By leveraging advanced parallel architectures, the proposed model optimizes feature learning and parameter efficiency, outperforming conventional deep learning (DL) frameworks. To evaluate its effectiveness, we compare our hybrid model with established transfer learning (TL) architectures, including VGG16, MobileNetV2, and ResNet50. A comprehensive class-wise performance analysis was conducted to assess the model’s robustness in detecting both masked and unmasked faces. Experimental results demonstrate that our hybrid architecture achieves superior performance, with an accuracy of 99.49%, precision of 98.95%, recall of 100%, and an F1-score of 99.48%. These results indicate a significant improvement over traditional TL models. The high accuracy and reliability of our model suggest its practical viability for real-time face mask detection in public health monitoring and disease prevention efforts. The proposed approach offers a powerful and effective solution for enforcing mask compliance, thereby contributing to global efforts in mitigating infectious disease transmission.