Deep Learning based Face Mask Detection

Yugansh Gaur, Aditya Agrawal, Pinki Nayak

Department of CSE, Dr. Akhilesh Das Gupta Institute of Technology and Management, INDIA

ABSTRACT

Severe acute respiratory syndrome corona virus 2 (SARS-CoV-2) pandemic is still continuously spreading all over the world. With the advent of the omicron variant, the situation is still under speculation in many parts of the world. Different studies have shown that usage of masks is the most significant way to avoid spreading of the virus from one person to another through air. Hence, it becomes crucial for people in public areas to wear masks. In this paper, the model trained and tested helps to identify individuals who are wearing a mask properly or not.

This could be done in a public place which can be monitored using cameras such as Closed-Circuit Television (CCTV) cameras. The model works on image processing and deep learning technologies.

The model was trained on a real-world dataset and successfully tested using live video streaming. One can improve the model's accuracy by experimenting with different variables, many people at varying distances and hyper characteristics and the frame's position. This study is beneficial in combating the spread of the virus and avoiding contact with the virus. The proposed model's superior performance makes it ideal for video surveillance systems.

Keywords: Deep Learning, Computer Vision, OpenCV, Tensorflow, Keras.