ENHANCED REAL-TIME DETECTION OF FACE MASK WITH ALARM SYSTEM USING MOBILENETV2

Abstract

To prevent the coronavirus from spreading, the government adopted measures such as wearing a face mask in public locations.  The researchers aimed to create a face detection system using the MobilenetV2 architecture that would identify a person’s faces and determine whether they were wearing a face mask. The built model will help to reduce the danger of viral transmission. In this study, face mask detection is achieved using a machine learning algorithm and the classification method using MobileNetV2. The steps for building the model are data gathering, data pre-processing, splitting the data, testing the model, and implementing the model. The built model can distinguish between those who are wearing a face mask (with no design patterns) and those who are not wearing it with a 96% accuracy. In terms of classification accuracy, the proposed model using MobileNetV2 outperformed the other models LeNet-5, AlexNet, and ResNet-50. If the detected person is labeled with “no mask”, the system generates an alarm sound. This research will be useful in combating virus spread and avoiding virus contact.