Publication Date: 2021/06/29
Abstract: The continuous spread of Corona virus has led to sustained increase in the mortality rate of many countries across the globe. WHO has made the use of masks mandatory in largely crowded areas which have reported Covid-19 cases. In order to curb the spread of virus and prevent cases, the detection of violators is highly desirable. We propose a model which highlights the use of deep learning approaches to identify people who do not wear mask. As most of the institutions, companies, industries, malls, hospitals, etc. have to start operating with few relaxations before this pandemic is completely erased, integrating face mask detection system with the existing surveillance systems at entry and exit points is highly recommended. The face mask detection model is built upon the MobileNetV2 architecture and detects face masks along with the percentage accuracy of wearing the mask properly in crowded scenes, both in images and real time streams. As the next step, the fresh concept of introducing facial identification of defaulters acts as a measure to keep the authorities informed of the people who are violating Covid-19 policies. Under the current Covid-19 lockdown time such system is definitely important to prevent the spread in many use cases. Airports, Hospitals, Offices will be some places which will benefit out of this system
Keywords: Mask Detection, Face Recognition, Database Connectivity, MobileNetV2, Computer Vision, COVID-19
DOI: No DOI Available
PDF: https://ijirst.demo4.arinfotech.co/assets/upload/files/IJISRT21JUN571.pdf
REFERENCES