Abstract—As the world experiences a pandemic caused by the spread of COVID-19, following health protocols like the proper wearing of face masks is essential to help prevent the spread of the virus as stressed by the World Health Organization (WHO). In the Philippines, manual monitoring of these measures may not be as efficient when handling large crowds. The authors were able to generate a real-time, scale-invariant face mask recognition model that also handles occlusions. It was trained using images from the wild to simulate real-world conditions. Existing state-of-the-art object detection frameworks such as MobileNet, RetinaNet, and YOLOv4 were implemented to detect and localize the proper use of face masks in public. To evaluate the performance of the models, the mAP, inference time, and precision curve were used. The performance of the model has resulted to as high as .94 mAP with an inference time of 73 ms.
Index Terms—Face mask detection, object detection, deep learning, occlusion, scale-invariance.
The authors are with the Don Mariano Marcos Memorial State University, Agoo, \Philippines (e-mail: lpfacun@student.dmmmsu.edu.ph, mjbaculo@dmmmsu.edu.ph, mlibao@student.dmmmsu.edu.ph, ceisma@student.dmmmsu.edu.ph, cfredeluces@student.dmmmsu.edu.ph, rgarlejo@student.dmmmsu.edu.ph, ridio@student.dmmmsu.edu.ph).
Cite: Louis Philippe B. Facun, Maria Jeseca C. Baculo, Marlon F. Libao, Ceazar M. Eisma, Christian B. Fredeluces, Rigzor A. Garlejo, and Raymart S. Idio, "Object Detection Frameworks for Real-Time, Scale-Invariant Face Mask Detection," International Journal of Future Computer and Communication vol. 11, no. 1, pp. 12-17, 2022.
Copyright © 2022 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).
Copyright © 2008-2024. International Journal of Future Computer and Communication. All rights reserved.
E-mail: ijfcc@ejournal.net