Facial Recognition Software Leads To Arrest

Share on FacebookTweet about this on TwitterShare on Google+

bigstock-Female-face-with-lines-from-a-45276856-compressed

In 1999, Neil Stammer was arrested in New Mexico on child sex abuse and kidnapping charges. Once released on bond, he forged a fake passport using the name “Kevin Hodges” and fled to Nepal, where he quietly hid out and posed as an English teacher.

The FBI’s investigation into Stammer’s whereabouts eventually went cold, and Stammer’s face remained on the FBI’s Most Wanted list for the last 15 years — until this month.

Last week, it was announced that Stammer was finally caught and extradited to the U.S., and law enforcement caught him by using a tool that we are sure to see more and more of in the future — facial recognition software.

Recently, the Diplomatic Security Service, which “protects US embassies and checks the validity of US visas and passports,” began using FBI wanted posters to test out the FBI’s latest facial recognition system, called the Next Generation Identification system (NGI). Eventually, they got a positive match for Stammer’s FBI wanted poster when the system came across the fake-Kevin Hodges passport photo, finally bringing the criminal to justice more than a decade later through advances in technology.

“He was very comfortable in Nepal,” said FBI special agent Russ Wilson. “My impression was that he never thought he would be discovered.”

The Electronic Frontier Foundation (EFF), a civil liberties group that has brought several lawsuits against the FBI over the years, reports that the FBI’s NGI database contained nearly 16 million images in 2013, and is projected to contain roughly 50 million images by 2015. The database also reportedly stores fingerprints, iris scans and palm prints.

It’s important to note that facial recognition software has been shown to be most effective when it is based on comparing images that have been taken under ideal circumstances — that is, pictures in which the person is looking straight into the camera, with their whole face properly exposed, just as you’d see in any drivers license or passport photo. The technology is significantly less efficient when it cannot clearly see a face that is squared up to the camera, for example as in surveillance camera footage, in which the camera is often attached to the ceiling and pointing down at an unusual angle or in which the camera has captured events at night.

We still have a ways to go before a surveillance camera can pick someone out of a crowd and identify them using facial recognition software. But situations like this one will ultimately prove key to fine-tuning the technology.

 

***UPDATE***

As of September 15, 2014, the FBI has fully rolled out their Next Generation Identification System  (NGIS). Not only is this program designed to expand biometric identification throughout the United States, but eventually the FBI plans to use the NGIS to replace their current fingerprint system. Until then, NGIS will work in concert with the fingerprint system.

The NGI system entails two databases. The first database, called Rap Back, enables law enforcement and related agencies the ability  to receive ongoing status notifications of any criminal history reported on any specific individual. This feature is designed to help law enforcement agencies track and remain updated of suspects under investigation.

The second database, Interstate Photo System (IPS), is the actual facial recognition program. IPS has image searching capability for photos associated with criminal backgrounds.

The FBI’s Next Generation Identification System and its databases have now been made available to over 18,000 law enforcement agencies as well as other criminal justice partners 24 hours a day, 365 days a year.

This entry was posted in Enhancement of audio and video and tagged , , , , , , , , , , , , , , , , , , . Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *