with the Office of Justice Programs' National Institute of Justice (NIJ)
Facial ID system unmasks identity, boosts arrests
The deputy can continue to keep an eye on the driver while the computer automatically downloads the image
Pulled over for running a red light, the driver tells the officer who stopped him that his name is John Smith and he must have left his wallet at home. Does the officer a) let him go with a warning, b) take him into the station for fingerprinting and further attempts at establishing his identity, or c) take his picture?
If the officer is a Pinellas County, Fla., sheriff’s deputy, then “c” is the correct answer. After plugging the easy-touse digital camera into the car’s laptop, the deputy can continue to keep an eye on the driver while the computer automatically downloads the image, opens the sheriff’s office facial recognition program, converts the image with a binary algorithm, runs a search of the county’s database and produces a gallery of potential matches, all in less than 30 seconds.
It might turn out that the driver’s name isn’t John Smith after all.
Using initial funding provided through the U.S. Department of Justice’s Office of Community Oriented Policing Services (COPS), Pinellas County has adapted a facial recognition system that has grown from a replacement for the county jail’s mug shot database into a partnership system that encompasses 14 of the state’s 67 counties and could well serve as a model for imilar systems in other states. The software was developed by Viisage of Massachusetts.
Pinellas County Capt. Jim Main explains that when the project started in 2001, the idea was to use a facial recognition algorithm with seven years’ worth of jail system digital images to help positively identify individuals who might be giving fictitious names or who could not provide identification. From its inception, staff photographed everyone brought into the ounty jail at the sally port and compared their images to those already in the database.
“Shortly after we went live, the patrol deputies pointed out that if they pull someone over who is playing the ‘name game’ nd doesn’t have a driver’s license, they have to decide if there is justification for bringing that person in for fingerprinting,” Main says. “They thought it would be great to be able to take an image on the street and get results back.”
Pinellas County began phasing that capability into patrol cars in 2004. By 2009, deputies made 496 arrests that could be directly attributed to identification made by the facial recognition technology and confirmed another 485 identities that did not require arrest, according to Systems Analyst Scott McCallum.
“The premise was to keep it short and simple for the deputy,” Main says. “We didn’t want them on the side of the street aking extensive clicks and opening windows, so we worked with the vendor to completely automate development. The county has since obtained additional funding to expand the system to 14 major metropolitan counties throughout Florida, and recently initiated a pilot project with the Florida Department of Motor Vehicles. The partnerships enable Pinellas County to search the other counties’ databases, and vice versa.
“The more images you get, the greater chance you have of making a match,” Main says.
Although the Florida system has expanded about as far as the licensing agreement will allow, other states and jurisdictions can purchase their own licenses from the same vendor and use them to establish their own compatible systems that use the same binary algorithm.
Main says that agencies in both South Carolina and West Virginia have expressed interest in setting up similar programs. When these other systems come online, they will be able to transmit images back and forth to Pinellas County and perform reciprocal searches for each other.
Even without compatible interfaces between other agencies, Pinellas County provides mutual aid and performs searches of its system for outside requests. For example, a recent request from the South Carolina Fusion Center resulted in a positive ID for a man who had been using numerous driver’s licenses with different aliases.