Face Mask Detection During COVID-19 Pandemic Using NASNetMobile and CNN Deep Learning Algorithms

The coronavirus (COVID-19) pandemic has led to a health, economic, and social crisis, resulting in many casualties and disrupting daily life, trade, and global movements. The most common transmission mode of infection is through respiratory droplets loaded with viruses. They are transmitted during s...

Full description

Saved in:

Bibliographic Details
Published in:	2022 4th International Conference on Current Research in Engineering and Science Applications (ICCRESA) pp. 137 - 142
Main Authors:	Abdullah, Mohammed AbdulSattar AbdulGhani, Ibrahim, Laheeb Mohammed
Format:	Conference Proceeding
Language:	English
Published:	IEEE 20-12-2022
Subjects:	Classification algorithms Convolution Neural Network (CNN) Corona Protection COVID-19 Data models Deep learning Mask Detection NASNetMobile Neural networks Pandemics Training
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	The coronavirus (COVID-19) pandemic has led to a health, economic, and social crisis, resulting in many casualties and disrupting daily life, trade, and global movements. The most common transmission mode of infection is through respiratory droplets loaded with viruses. They are transmitted during speaking, coughing, or exhaling when people are close to each other and without protection covering the face. Wearing a face mask is the way to reduce COVID-19 transmission from one person to another in public places praised by the World Health Organization (WHO); manually knowing people wearing a face mask is very difficult, especially in crowded places. This paper aims to build a system based on NASNetMobile and Convolution Neural Network (CNN) Deep Learning Algorithms to capture frames from the webcam in realtime, detect faces, and then classify the face whether it is wearing a mask or not. The system is built starting with training dataset collection, preprocessing dataset, classification model building, evaluation, testing, and model implementation. The accuracy of the model built by the Convolutional Neural Network reached 99.3% when evaluated on the test data, while the model built by the NASNetMobile model architecture reached 99.76%. This system helps monitor crowded places and alerts cases of violations of the laws on wearing a face mask.
DOI:	10.1109/ICCRESA57091.2022.10352501