The Federal Bureau of Investigation (FBI) is betting big on facial recognition as the agency begins to roll out its long-awaited $1 billion Next Generation Identification (NGI) system. The program combines iris scans, DNA analysis and voice recognition alongside the ability to spot people based on their faces in a crowd.
The FBI has already began to test the program in select markets with a full nationwide launch expected by 2014.
The bureau plans to use the software for both tracking known criminals and for picking potential criminals out of faces in a crowd.
Using the new system images of a person of interest can be taken from security cameras and publicly uploaded photos on the internet. Those stills and photographs are then compared against a national repository of images that have been uploaded into the facial recognition program by FBI workers.
Privacy advocates in the meantime are worried that people with no criminal record will be entered into a federal database without their permission or be subject to unwarranted surveillance.
While the FBI has not explained the algorithm it is using for the facial recognition program experts believe it is very accurate, especially in controlled situations when faces are being pulled from police mugshots and passport photos.
According to New Scientist:
“Tests in 2010 showed that the best algorithms can pick someone out in a pool of 1.6 million mugshots 92 per cent of the time.”
Already joining in the program are state issuers of driver’s licenses which are providing the FBI with posed photos.