Real-Time Face Mask Detection Using Deep Learning on Embedded Systems

Date of Award

12-2021

Document Type

Thesis

Degree Name

Master of Science in Data Science

First Advisor

Patricia Angela R. Abu

Abstract

Coronavirus disease (COVID-19) is an infectious disease, which is caused by coronavirus 2 (SARS-CoV-2) that was identified in December 2019 in Wuhan, China. It is a pandemic that causes respiratory disorder and is transmitted through sneezing droplets of infected individuals. According to the World Health Organization Weekly epidemiological update [48], as of the 1st of June there are 169.6 million infected people and over 3.53 mil- lion deaths globally. This pandemic not only affects our health but also affects our lively hood. In the absence of specific treatment or a vaccine, non-pharmaceutical interventions (NPI) form the backbone of the response to the COVID-19 pandemic. These NPI includes physical distancing, regu- lar hand washing, and wearing a face mask. This study aims to help with the monitoring of these NPIs specifically wearing face masks using deep learning. This study implements face-mask detection and recognition sys- tem on embedded systems. It automatically detects and recognizes if a per- son is wearing a medically approved face mask, a non-medically approved face mask, or not wearing a mask. It uses transfer learning to train Con- volutional Neural Network (CNN) models to detect and classify face masks. The models are pruned and quantized to optimize them for embedded sys- tem implementation. From the results, pruned Mobilenet and Mobilenetv2 performed the best during implementation, with up to 4∼ 5 frames per sec- ond.

Link to Full Text

Share

COinS