Light-Weight Face Mask Detector

People’s lives have been severely disrupted recently due to the COVID-19 outbreak’s fast worldwide proliferation and transmission. An option for controlling the epidemic is to make individuals wear face masks in public. For such regulation, automatic and effective face detection systems are required. A facial mask recognition model for real-time video-recorded streaming is provided in this research, which categorizes the pictures as (with mask) or (without mask). A dataset from Kaggle was used to develop and assess the model. The suggested system is computationally more precise, efficient and lightweight when compared to other systems like VGG-16, DenseNet-121, and Inception-V3 which helped the developed model meet low end PC system requirements. The collected data set contains exactly 12,000 images and has a 98.1% performance training accuracy and a validation accuracy of 98.2%, which is achieved by using MobileNetV2.