Week 9: Project Literature Review & Methodology

LITERATURE REVIEW

A. The Importance of Mask during COVID-19

In order to slow the spread of COVID-19, limiting physical distancing and using other measurements to decrease the transmission probability are necessary. According to J. Howard et al., if face masks were used by all household members prior to symptoms occurring, it has a 79% effective rate in preventing transmission. Besides, the states with mask mandates to require residents to wear a mask have a lower daily growth rate than those without. As a mask provides a physical barrier that reduces the spread of respiratory droplets and particles into the air, properly wearing a mask can protect the safety of everyone [1].

Figure 1: The efficiency of mask in blocking the virus

B. The Implementation of Mask Detection Robot

Currently, mask and temperature detection robots have shown their important role in responding to the COVID-19 pandemic. The UBTech had developed some robots which can measure body temperature when people entered a facility [2]. This robot serves as the first line of defense to prevent person-to-person infection at the hospital. The pepper robot which was built by SoftBank is able to scan the faces of up to five people in a group simultaneously to check if they are wearing a mask, without the need for an internet connection [3]. 

Figure 2: Pepper Robot performs face mask detection

Along with this, the SMP security robot has the ability to detect a person without a mask at a short distance from 4 to 24 feet [4]. Besides, a four-leg robot has been used in public parks to remind people of keeping a safe distance and wearing masks in Singapore [5]. However, when working in the subway, these robots have common problems which include having over-sized bodies, consuming too much power, unable to work in narrow pathways. To solve these problems, this project aims to build a face mask detection robot with a small size and low-cost. 

Figure 3: The security robot developed by SMP

Methodology

To implement the face mask detection technique, we plan to build the detector using OpenCV, Keras/TensorFlow, and Deep Learning following the tutorial written by Adrian Rosebrock [8].The development of the detector is separated into two phases. In the first phase, we will train the detector using the dataset from the tutorial. Then, we will deploy the trained detector and classify each face as withMask and withoutMask. We will later combine the Roomba driving control and the detector together.

Figure 4: The development process of the face mask detector

References

[1] J. Howard et al., “An evidence review of face masks against covid-19,” Proceedings of the National Academy of Sciences, vol. 118, no. 4, p. e2014564118, 2021.

[2] E. Guizzo and R. Klett, “How Robots Became Essential Workers in the COVID-19 Response.” [Online]. Available: https://spectrum.ieee.org/robotics/medical-robots/how-robots- became-essential-workers-in-the-covid19-response

[3] “New Feature: Pepper Mask Detection,” Sep. 2020. [Online]. Avail- able: https://www.softbankrobotics.com/emea/en/blog/news-trends/new- feature-pepper-mask-detection

[4] “Face mask detection robot with a voice warning of a fine for not wearing it in the public area,” Nov 2020. [Online]. Available: https://smprobotics.com/usa/face-mask-detection-robot/

[5] Y. Li, “Robot dog enforcing social distancing in Singapore(1/4),” May 202AD. [Online]. Available: http://www.ecns.cn/hd/2020-05-15/detail- ifzwknkv0960244.shtml

[6] M. Loey et al., “A hybrid deep transfer learning model with machine learning methods for face mask detection in the era of the covid-19 pandemic,” Measurement,, vol. 167, p. 508288, 2021.

[7] G. J. Chowdary et al., “Face mask detection using transfer learning of inceptionv3,” in Big Data Analytics, vol. 12581, pp. 81–90, 2020.

[8] A. Rosebrock, “COVID-19: Face Mask Detector with OpenCV, Keras/TensorFlow, and Deep Learning,” May 2020. [Online]. Avail- able: https://www.pyimagesearch.com/2020/05/04/covid-19-face-mask- detector-with-opencv-keras-tensorflow-and-deep-learning/