Abstract
With the global spreading of the COVID-19 pandemic, the face-covering regulation has become a common disease control policy issued by the public authorities nationwide in the United States and the rest of the world. Many public and private service providers require people to wear an appropriate mask correctly; however, it is still challenging to monitor whether individuals practice such a policy or not. In the artificial intelligence era, deep learning models with computer vision technologies can be applied as an efficient solution for this challenge in the context of the COVID-19 pandemic. In this chapter, a deep transfer learning-based face mask detection model has been proposed with the testing procedure for multiple purposes, such as identifying whether individuals wear face masks or not, recognizing which type of the face-covering a person wears, and detecting whether individuals wear masks correctly or not, with a real-time human-computer interaction representation.
TopIntroduction
In today’s age, COVID-19 becomes a crucial public health issue that affects people’s quality of life, leading to huge impacts to the social and economic system. The new coronavirus spreads easily and widely in crowd and close contact due to its high transmission rate, and becomes one of the most significant challenges around the world. Advised by health organizations and epidemiologists, the usage of personal protective equipment (PPE) has been recommended in crowded and closed environments (Shen et al., 2021; Garrigou et al., 2020). The common types of PPE, such as gloves, eye protections, gowns, and face masks, can minimize exposure to the virus, thereby reducing the risk of the COVID-19 infection (Howell et al., 2021). Public authorities have begun working on making new rules by forcing individuals to follow the disease control policies, including practicing social distancing and facial covering. Wearing a mask becomes the requirement in public area in order to reduce the infection and spreading rate of COVID-19 since it can be transmitted via airdrops. As normal people are forced by the face covering regulations across the globe, it is challenging to monitor whether individuals follow such rules or not, especially in the high density of the population. In order to win the battle against the pandemic, guidance and surveillance in the crowd have been deployed by government officials to ensure that face covering regulations are applied with computer vision technologies and object detection algorithms (Wu et al., 2020). Such attempted applications can be implemented by integrating surveillance systems and artificial intelligence (AI) techniques.
Despite the huge impact on the society, the COVID-19 pandemic has given rise to opportunities in terms of research cooperations and industrial applications. Artificial intelligence along with computer vision technologies provide cutting-edge solutions to fight against the virus in many forms. As introduced in previous chapters, machine learning and deep learning allow research scientists and policymakers to build the early warning mechanism for the pandemic outbreaks and to monitor human behaviors in regards of disease control measures and regulations. The computer-based face covering detection, as one of the innovated approaches in the social control system during the pandemic, has been applied in the process of monitoring large groups of people. A novel AI-based prototype has been applied in the surveillance cameras of the Paris Metro System in order to monitor whether riders wear face masks or not (Fouquet, 2020). A face mask detection system, developed by LeewayHertz, has been proposed by using existing IP cameras and CCTV surveillance cameras combined with computer vision to monitor people without face masks (Pispati et al., 2020). Executed by the National Institute of Standard and Technology (NIST), a study has been proposed under the Ongoing Face Recognition Vendor Test (FRVT) documents accuracy of algorithms to detect masked individuals in the United States, which is the first report on evaluating the performance of face recognition models on protective face masks in COVID-19 (Ngan et al., 2020). The national first facial recognition tool that can identify citizens when they are wearing a face mask, developed by a Chinese facial recognition company, has been developed for assisting disease control purposes in China, which is more technologically advanced because it can extract personal information from mask-wearers (Pollard, 2020). Although ethical issues of the usage of facial recognition for surveillance systems in public have been argued for a lone time (Brey, 2004), its application focusing on the face mask detection is significant and demanded. Reported by DatakaLab, the purpose of using face mask detection algorithms is not to arrest individuals who violate the rule but to provide scientific suggestions that can help governments to reduce the risk from COVID-19 (Faizah, et al., 2021).
Key Terms in this Chapter
Face Mask Detection: A computer-based monitor that detects whether individuals are wearing a mask or not.
Deep Transfer Learning: A optimized transfer learning method that can decrease the distribution difference between source and target domains.
Human-Computer Interaction: A study field that focused on designing the interactive tools through applying computer technologies.
Deep Learning: A broad family of machine learning models based on neural networks. Typical deep learning models are deep neural networks, convolutional neural networks, recurrent neural networks, deep belief networks, and deep reinforcement learning.
Computer Vision: An automation technology that makes computers to gain high-level understanding from images and videos throughout acquiring, processing, analyzing, and recognizing digital data by transforming visual images into numerical or symbolic information.
Object Detection: A computer vision technique that can recognize objects from image or video.
Image Recognition: One of the most classical issues in computer vision, image processing, and object detection, which deals with determine whether or not an image contains specific objects, patterns, or features.
Convolutional Neural Network: A typical deep learning model that is commonly used to image classification, object detection, natural language procession, and predictive analysis. Such a network structure is a regularized version of fully connected networks, which belong to the class of artificial neural network.