The rampant spread of COVID-19 around the globe continues to challenge public health officials and governments alike to find ways to manage the spread of the disease until an effective vaccine can be developed. This challenge has led to new and novel approaches, particularly in the use of technology. One of the most recent technology applications to control the spread of the virus is the use of security cameras combined with facial recognition software.
Facial recognition is part of a computer software category that uses video content analysis (VCA) technology. VCA technology uses machine learning and artificial intelligence to detect objects within a video image and then extract, identify, classify, and index them for a broad range of applications.
As schools and businesses reopen, contact tracing has become an essential tool in preventing the spread of COVID-19. However, not surprisingly, people infected with the virus often struggle to remember everyone they have come in contact with in the previous two weeks, which reduces the effectiveness of contact tracing.
Video security systems
Instead of relying on human memory, schools and businesses that have video security systems can use facial recognition to quantify people's proximity across time and location. They can track where on the premises a student or employee has been and identify any other people that person has been in contact with while in the building.
This technology helps avoid having to close an entire school or business when an individual tests positive for the virus by providing more precise information about what areas need to be sanitized and who may need to be quarantined.
Safe and responsible reopening
VCA technology and video security systems can also be used together to help ensure safe and responsible reopening
In addition to contact tracing, VCA technology and video security systems can also be used together to help ensure safe and responsible reopening during the current pandemic in other ways, too:
Improving compliance with face mask requirements -- Providing the ability to search for people with or without a mask, facial recognition can be used to detect face mask violations in real-time and alert those responsible for ensuring compliance.
Limiting occupancy to ensure proper social distancing -- Video management systems software can be encoded with lower occupancy thresholds and rules to count the number of people entering and exiting a building or an area within it and send alerts when the occupancy thresholds are met. This allows security staff to ensure proper social distancing and provides a better understanding of where social distancing may be more challenging.
Here is one example of how a business might use facial recognition for contact tracing. When an employee self-discloses that he/she has tested positive for COVID-19, the employer can upload a digital image of that employee into its VCA system to conduct a filtered search through its video footage for the last 2-3 weeks for any face matches for that employee.
When matches are identified, the operator can review the video for each match to identify where in the facility the employee has been and who the employee may have come in contact with. The employer can then notify those individuals that they may have been exposed to the virus and recommend or require that they self-quarantine for the recommended 14 days.
It is incumbent on the employer and required by the federal Health Insurance Portability and Accountability Act (HIPAA) to protect the individual’s identity when notifying the people he/she has interacted with.
Critics argue that using video surveillance with facial recognition in hospitals and public spaces creates privacy issues
Research has found that facial recognition is not as accurate as people may think. In an analysis of the use of facial recognition technology in law enforcement, Cardiff University found thousands of false-positive matches. Concerning facial recognition algorithms' accuracy, the National Institute of Standards and Technology (NIST) defines a false positive as two different individuals incorrectly identified as the same person. A false negative means that the software failed to match two images of the same person.
The fact is, facial recognition technology has been controversial since its development. While facial recognition has been used to locate missing children and has improved the security at airports against terrorism, critics argue that using video surveillance with facial recognition in hospitals and public spaces creates privacy issues. In contrast, others point to concerns that inaccurate results can lead to false arrest problems when used in law enforcement.
Given the strong feelings that the use of this technology can elicit, any organization considering using facial recognition should be prepared to address them openly.
The Brookings Institute has developed several recommendations to help protect people from the potential problems facial recognition software can pose. These recommendations were developed prior to the emergence of COVID-19.
However, while many of them will take years to implement, there are two that, in the short term, can do much to help ensure the responsible use of facial recognition in preventing the spread of the disease:
Limit the Data Storage Time -- This is a reform that could go far in mitigating privacy concerns and fears around the misuse of data for purposes other than that for which it was originally collected. Data collected for contact tracing will no longer be relevant after the pandemic is over; therefore, there is no need to retain it beyond that. Defining limits on how long such videos will be retained will instill confidence that their images are used only for beneficial purposes and only for a specific period.
Provide Clear Notification in Public Areas that Facial Recognition is Being Used and Why -- This would allow those who object to avoid those areas. While it would seem that everyone who wants to do whatever they can to help control the spread of COVID-19, some may not agree that facial recognition is an appropriate way to do that. Transparency concerning the use of these technologies in public spaces is therefore very important.