Lynn Cozart
Lynn Cozartwas went missing two months after he was convicted of sexually molesting three children in 1996Police handout

A new facial recognition system being used by the FBI has managed to track down and capture a convicted paedophile on the Most Wanted List who disappeared 19 years ago.

Lynn Cozart, 63, from Beaver County, Pennsylvania was convicted for sexually assaulting his three children in February 1996, and he went missing just before his sentencing hearing two months later.

He managed to leave the state and stay under the radar for almost two decades, until the FBI used the $1bn (£650m, €900m) Next Generation Identification (NGI) system built by Lockheed Martin to track him down.

The FBI submitted Cozart's mug shot to the NGI system, which used facial recognition to capture biometric data of his face.

The system then sought matches from other databases belonging to state agencies across the US, from government-issued licence databases to any form of government-collected data of scanned facial images taken from video or photos.

The system spotted a match amongst driving licence photos held by Arkansas' motor vehicle department, and from there, the fugitive was tracked down to yet another state – Muskogee in Oklahoma, more than 1,160 miles away from Pennsylvania.

Cozart was found working in Walmart under an assumed name, David Stone, and was apprehended by the town's police officers.

What state-of-the-art identification systems can do

NGI was developed over multiple years to replace the FBI's existing Automated Fingerprint Identification System (IAFIS). It was rolled out in September 2014, containing almost 125 million fingerprints from both criminals and civilians, as well as 24 million mugshots.

The new system includes advanced fingerprint identification technology (AFIT) that improves the accuracy of matching fingerprints up to more than 99.6% from 92% with the old system, and the number of manual fingerprint reviews required by police officers is cut by 90%.

Then there is the repository for individuals of special concern (RISC), which can rapidly search a captured face from videos or photos against a database of wanted persons, sex offenders, known or appropriately suspected terrorists and other people on the persons of special interest list.

Millions of palm prints are also stored on the national palm print system, which is now part of NGI, and it is possible for the system to scan and analyse tattoos, scars and even irises for a match.

The use of facial recognition software by law enforcement has not been well-received by privacy groups, and the Electronic Frontier Foundation in particular sued the FBI in 2013 to find out more information.

"NGI will result in a massive expansion of government data collection for both criminal and noncriminal purposes," EFF staff attorney Jennifer Lynch said at the time.

"Biometrics programs present critical threats to civil liberties and privacy. Face-recognition technology is among the most alarming new developments, because Americans cannot easily take precautions against the covert, remote, and mass capture of their images."

In May, Lockheed Martin revealed that it has also worked on technology for the NSA that can identify individuals by their finger swipes on a smartphone.