In Pennsylvania, authorized law enforcement can access the state’s Justice Network (JNET) to access critical information for investigations. The JNET Facial Recognition System allows about 500 of JNET’s 40,000 users – individuals from local, state and federal law enforcement – to narrow down a search for a suspect by comparing images from sources like surveillance footage and social media sites against a statewide criminal database containing 3.5 million photos, according to JNET officials.
But since JNET’s Facial Recognition System could only compare photos against images of previously arrested or convicted criminals listed in the statewide database, JNET needed a new method for finding suspects unknown to law enforcement. To broaden its image comparison capabilities, in May, JNET’s facial recognition technology was integrated with the Pennsylvania Department of Transportation (PennDOT)’s facial recognition system. PennDOT stores 36 million driver’s license and identification photo images in a photo repository.
PennDOT maintains its own facial recognition system due to concerns over the issuance of duplicate or fraudulent licenses. Luci Stone, a JNET criminal biometrics specialist, said because the system is only used for those specific criminal investigation purposes, the technology is not in use for image comparisons on the general public.
“They’re looking to make sure no one steals someone’s identity or tries to get a couple different licenses,” said Stone.
And the same goes for the newly integrated system. Stone said law enforcement must adhere to citizen privacy guidelines and can only use the integrated facial recognition technology when pursuing a suspect during an investigation.
PennDOT’s facial recognition system was previously only available to a small group of users including the Pennsylvania State Police and the Office of Attorney General. Once the facial recognition systems were integrated, JNET officials report that the single interface still upholds law enforcement policies and citizen privacy.
“There came an opportunity recently for us to integrate the two systems and really do two things at once,” said Dave Naisby, JNET’s executive director. “The first was reduce costs for the Commonwealth, but then the second was to allow our users who had access to both systems the opportunity to search both systems through one user interface.”
According to Stone, only investigators and detectives who have been trained to use the facial recognition system can search across both the criminal and PennDOT databases by using three unique search algorithms.
Now that both systems have been integrated onto a single interface, Stone said results have been positive. Members of law enforcement have gotten leads using the integrated system that have led to suspects being caught.
According to JNET, the project cost the state $75,000 to develop, test and deploy.
