This archived news story is available only for your personal, non-commercial use. Information in the story may be outdated or superseded by additional information. Reading or replaying the story in its archived form does not constitute a republication of the story.
SANDY — A face can hold the vital clues detectives need to track down those suspected of terrorist acts and crimes of all kinds, and software specially designed to potentially spot that face is already paying dividends in Utah and in other states across the country.
Utah Department of Public Safety Maj. Brian Redd said the facial recognition system installed at the Statewide Information & Analysis Center in Sandy has been in place since 2009 and is now used in roughly a thousand checks per year.
“We’ve helped law enforcement agencies solve frauds and home invasions and bank robberies,” Redd said, also noting the technology was once even used in a case to put a name to an unidentified man who was discovered badly injured on the side of a road. “It’s very beneficial. It’s been a great tool for us.”
Redd said the technology isn’t like the movies, where people can be identified off of live surveillance video feeds. Instead, he said law enforcement agencies will bring a photo to SIAC in conjunction with a criminal investigation, in order to compare the image against jail booking photos, prison pictures and photographs housed in the state’s driver license database.
The fusion center does not have immediate access to federal database photos and requires assistance from the Federal Bureau of Investigation for those, Redd said.
How it works
SIAC director Steve Hewitt said the facial recognition system runs photos against 17 different algorithms and comes up with percentage matches.
“The initial search is set fairly high in terms of the percentage match, but sometimes because of lighting, because of angles, things of that nature, we have to set that back to bring in some additional images to further evaluate,” Hewitt said.
Hewitt said analysts have to use their training and ability to weed out false positives and identify the correct individual.
“There’s natural age progression issues that you have to factor into this, as well as if an individual were trying to hide his identity or change his identity,” Hewitt explained. “It’s about the analysts’ evaluation and analysis of not only the image, but any additional data that is available to them as well as providing that back to the investigator so that they can further conduct their investigation and make some decisions.”
Hewitt said the system has significantly reduced the amount of time it takes for detectives to identify suspects in identity fraud-type cases, which he said would have been much more challenging 15 years ago.
He pointed to one recent case where the U.S. State Department had been investigating a man the department suspected of passport fraud. Analysts were able to identify two additional identities of the man in the system, and they provided that information back to state department officials for their investigation, according to Hewitt.
“It’s very important and a very important tool because we are concerned about essential individuals from outside the country coming into the United States and conducting crime or terrorism,” Hewitt said.
Redd said the Utah Driver License Division also uses facial recognition software every time somebody takes a new photo for a driver’s license.
“We’re trying to ensure the driver’s license database doesn’t have fraud,” Redd explained. “When you come in to get a photo taken, it’s run in the facial (recognition) system to make sure your photo and your identity is not being used by someone else.”
Redd says the division does not encounter many “hits.”
“(That) tells us we feel pretty confident that it’s got integrity – the driver’s license database,” Redd said.
Facial Recognition and Privacy
Both Redd and Hewitt underscored the care and caution with which the system is used and handled.
Redd said when photos are run against the available databases, there’s no personal information included, and in “almost all cases” what is returned to law enforcement agencies also does not include personal identifying information.
Redd said there are no unnecessary searches, and searches are only tied to active criminal investigations where there is a case number and a photo.
“We support criminal investigations; we don’t conduct criminal investigations,” Redd said.
Redd said those behind the fusion center speak regularly with legislators and legislative leadership, as well as civil liberties watchdogs and citizens when they have questions.
“We go to great lengths to make sure this is a transparent operation and that we do protect the privacy of citizens,” Redd said.
Redd said with those precautions in place, the facial recognition software is an effective investigative tool.
“It’s helped us solve crimes; crimes that we would probably not have otherwise solved,” Redd said. “We use it very limited, we use it when there’s a criminal investigation. We use it when it’s important, and we solve crimes.”