Facial recognition 'proven' as airport crowd filter
Facial recognition cameras have been proven as a means of spotting wanted people in crowds, claimed LogicaCMG.
However, the proof was limited and the owners of public venues have proven reluctant to pay for the technology because they believe their business is entertainment rather than security.
LogicaCMG achieved an average 85 per cent success during a pilot of facial recognition technology in an unnamed international airport rate in singling people out after matching them against a database of 1,000 wanted people.
Tim Best, director of global identity solutions at LogicaCMG, said: "It's the first time anyone has got 85 per cent match with a live video feed.
The best results were achieved in corridors where there was a relatively controlled flow of well-behaved people. In less favourable conditions like on a stairwell, where people where watching their feet, the hit rate was something like 65 per cent. The airport had wanted to see a 75 per cent success rate. The idea is that those pulled up by the system, including the 20 to 40 per cent of people wrongfully identified, would be pulled over by customs officials.
Best said he would have to do a full-scale pilot to determine whether the technology could deal with the brimming crowds of an airport concourse.
Urs Schmeid, director of biometric solutions at Unisys, said such results were of limited interest to commercial customers because they didn't think they ought to pay out for security that might be the responsibility of the state.
"We've had [facial recognition] pilots, but the interest isn't as high as we thought. It's a question of spending money.®