MobileNet Mask: A Multi-phase Face Mask Detection Model to Prevent Person-To-Person Transmission of SARS-CoV-2
Medical researchers around the globe provide evidence that COVID-19 pandemic diseases transmitted through droplets and respirators of respiratory aerosols and wearing a face mask is an efficient infection control recommendation process. In addition, many
- PDF / 788,427 Bytes
- 11 Pages / 439.37 x 666.142 pts Page_size
- 60 Downloads / 227 Views
Abstract Medical researchers around the globe provide evidence that COVID19 pandemic diseases transmitted through droplets and respirators of respiratory aerosols and wearing a face mask is an efficient infection control recommendation process. In addition, many public and private service providers demand that consumers use the service only if they wear masks properly. However, a few research studies have been found on face mask detection based on the technology of Artificial Intelligence (AI) and Image Processing. In this article, we propose, MobileNet Mask, which is a deep learning-based multi-phase face mask detection model for preventing human transmission of SARS-CoV-2. Two different face mask datasets along with more than 5,200 images have been utilized to train and test the model for detecting with and without a face mask from the images and video stream. Experiment results show that with 770 validation samples MobileNet Mask achieves an accuracy of ~ 93% whereas with 276 validation samples it attains an accuracy of nearly ~ 100%. Lastly, we also discuss the possibility of implementing our proposed MobileNet Mask model on light-weighted computing devices such as mobile or embedded devices. Besides, this proposed model also introduces frontier technologies to support the efforts of government and public health guidelines with anticipation of implementing mandatory face mask regulations all over the world. Keywords SARS-CoV-2 · Face mask · COVID-19 · MobileNetV2 · Deep learning
S. K. Dey (B) Dhaka International University, 66 Green Road, Dhaka 1205, Bangladesh e-mail: [email protected] A. Howlader Patuakhali Science and Technology University, Dumki 8602, Bangladesh e-mail: [email protected] C. Deb Cyber City Tower 2, Pune, Maharashtra, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 M. S. Kaiser et al. (eds.), Proceedings of International Conference on Trends in Computational and Cognitive Engineering, Advances in Intelligent Systems and Computing 1309, https://doi.org/10.1007/978-981-33-4673-4_49
603
604
S. K. Dey et al.
1 Introduction A few people used to wear face masks in order to protect them from air pollution. During COVID-19 epidemic, almost every healthcare professional and scientist recommended wearing face masks in public places. COVID-19 is arising issue from respiratory virus infections caused that emerging by Severe Acute Respiratory Syndrome coronavirus 2 (SARS-CoV-2). SARS-CoV-2 transmission caused through droplets that ejected from person’s speaking, coughing, or sneezing with common droplet size 5–10 µm but aerosol emission and droplets emission increased when human speech and shout loudly [1]. Many laboratories setup research evidence that surgical masks were found to filter better than cloth mask filtered because large droplets evaporate and turn into minor aerosol particles such as 3 to fivefold smaller. Infected aerosol particles produced by patients coughing where this airborne virus aerosol atoms c
Data Loading...
