Facial recognition technology picks "suspects" out of an airport crowd
LogicaCMG's facial recognition trial has 85 per cent success rate at a major airport, even when volunteers disguised themselves.
A facial recognition trial has had an 85 per cent success rate at picking faces out of a crowd at one of the world's leading airports, the IT and business services firm LogicaCMG has announced.
LogicaCMG ran the trial in an undisclosed airport to see if individuals could be successfully identified from a distance of several metres over CCTV technology.
Biometric identification has long been possible, but identifying a person at a distance is much more difficult, with special technology or without. CCTV operators scanning footage for specific faces can experience monitor blindness after just 20 minutes, LogicaCMG said in a statement.
"Using biometric technology, such as finger printing or iris recognition, to verify someone's identity has been possible for some time," said Tim Best, director of global identity solutions at LogicaCMG. "The actual identification of a person is significantly harder which is why it has had to be done manually for so long, despite the limitations."
During the trial, volunteers photographs were put on a database's suspect list. While the volunteers wandered the airport and mingled with real passengers, a biometric identification algorithm was used to analyse the faces in the crowds for a match.
Thousands of other photographs were also included on the database to test the system for false identifications.
The trial had a success rate over 85 per cent, LogicaCMG said, even when the volunteer "suspects" changed their facial hair or wore glasses.
While the technology has a clear use for airport and border control security, Best said it could also be used in other areas. Similar technology is already used by a number of commercial organisations in the US such as casinos in Las Vegas, which use facial recognition via video to spot cheats and other banned patrons in their facilities.
"The advantages of being able to identify non-compliant people from a distance are clear for the border control industry, however we envisage this technology being extended and applied to other areas, like helping to reduce hooliganism at sports events or persistent shop-lifting in retail outlets," said Best.
IT PRO previously reported that Southampton University is researching the use of gait recognition in airports.
Unleashing the power of AI initiatives with the right infrastructure
What key infrastructure requirements are needed to implement AI effectively?Download now
Achieve today. Plan tomorrow. Making the hybrid multi-cloud journey
A Veritas webinar on implementing a hybrid multi-cloud strategyDownload now
A buyer’s guide for cloud-based phone solutions
Finding the right phone system for your modern businessDownload now
The workers' experience report
How technology can spark motivation, enhance productivity and strengthen securityDownload now