Facial recognition AI can identify mask-wearers, US research finds
Facial recognition technology can identify people wearing face masks, tests conducted by the Department of Homeland Security (DHS) in the US suggest. The DHS hopes to reduce the need for people to remove their face masks during ID checks in facilities such as airports.
The testing was conducted as part of the 2020 Biometric Rally, held in Maryland last autumn by the Science and Technology Directorate (S&T) within DHS. The focus was on exploring if biometric systems could “detect and recognise” mask-wearing travellers.
“In many cases, photo IDs are the best or only way we currently have to verify our identity. Current processes and technologies generally require the face visible. Asking travellers to temporarily remove their mask and put it back on is not an ideal solution,” wrote Arun Vemury, director of S&T’s Biometric and Identity Technology Center, in a feature in December 2020.
Over 10 days, the DHS said, 582 people from 60 countries participated in the research, which tested 60 facial recognition configurations drawn from six face and/or iris systems and 10 matching algorithms.
The early results, released this week, show that on average the systems identified about 77% of people when wearing masks. The best-performing system identified about 96% of people with masks but, the DHS notes, “performance can vary greatly between systems”.
This was a clear reduction in performance when compared with mask-free testing. On average, when people did not wear masks, all systems identified 96% of people, with the best identifying 100% of participants.
“This isn’t a perfect 100% solution,” said Vemury, “but it may reduce risks for many travellers, as well as the frontline staff working in airports, who no longer have to ask all travellers to remove masks.”
The final test results from the 2020 Biometric Technology Rally will be released in the coming weeks, according to DHS.
About Kate Hodge
Kate is a journalist and editor, holding roles at both the Guardian and the Financial Times. She specialised in education and combines writing, commissioning and editing with social media and audience engagement. If you have any ideas you would like to pitch, or suggestions to improve the website, feel free to email her on [email protected].
Related Posts
