FaceShroud Advanced Real-Time Mask Detection using Deep Learning

COVID-19, an infectious disease caused by a novel coronavirus, manifests with a range of respiratory symptoms, from mild to severe, potentially worsening without prompt medical intervention. Individuals with pre-existing health conditions, such as cardiovascular disease, diabetes, chronic respiratory ailments, and cancer, face a higher risk of developing severe illness. Comprehensive knowledge about COVID-19, its transmission, and associated illnesses is crucial for prevention and containment efforts. Wearing masks is a key measure in controlling the virus spread. Implementing classical deep learning techniques for real-time face mask detection aids in identifying individuals without masks, thus contributing to virus transmission control. The face mask detection dataset comprises three classes: mask_worn, mask_worn_incorrectly, and no_mask. OpenCV facilitates real-time face detection from webcam streams. Deep learning architectures such as InceptionNet and XceptionNet, employing transfer learning, are utilized for classification. InceptionNet achieves training and validation accuracies of 93% and 90%, respectively, while XceptionNet achieves a training accuracy of 98% and a validation accuracy of 90.74%.