FBI delivers on $1.1B biometrics program

Listen to Jason's story on the Federal Drive

wfedstaff | June 4, 2015 8:26 pm

The FBI added a host of new capabilities for federal and local law enforcement officials to improve how they identify suspects. The Next Generation Identification System replaced a 15-year-old fingerprint system with advanced biometrics.

But it wasn’t just the facial recognition or the move from two fingerprints to 10 fingerprints or the huge improvements around identifying latent fingerprints. The FBI actually delivered this $1.1 billion program on-time, on-budget and within expected capabilities, and turned off the 1990s-era Integrated Automated Fingerprint Identification System (IAFIS).

“It was a quite a feat. It was actually a 10-year program with six years of contract work. Increments we decided were the best way to deliver functionality each year. We can make sure we have something out there,” said Jim Preaskorn, the program manager for the FBI’s next generation identification system, in an interview with Federal News Radio after the AFCEA Bethesda’s 8th Annual Governmentwide Initiatives Excellence Awards in Washington. “As for the cost and schedule, it was a team effort. It was a $1.1 billion program, so we had a whole program office and that philosophy of earned value, schedule containment and quality as well as scope. Think of it as a three-legged stool, quality, cost and schedule, saw one off, you have to do something to the others.”

The FBI hired Lockheed Martin in 2008 under a 10-year contract to implement NGI to replace IAFIS, which the bureau implemented in 1999.

Advertisement

The FBI and Lockheed began implementing new capabilities starting in 2011, with increment four going live in September 2014 to meet full operational capability.

It was that reason why AFCEA Bethesda honored Preaskorn and the FBI with the Excellence in Mission Award. In all, the organization honored federal and state IT executives in nine categories.

“You are being honored here today because you naturally aspire to and exhibit these ideals of excellence. You are being honored here today because you passionately and instinctively employ the attributes we have heard extolled,” said Richard McKinney, the Transportation Department chief information officer, who gave the keynote speech at the awards ceremony. “You understand their meaning intuitively without needing a reminder from me or anyone else because you practice them every day. You work hard as a matter of habit. Your virtue and your excellence are the just rewards of that hard work and the cause for this celebration today.”

Accuracy and speed

The FBI’s hard work translated into significant more capabilities for federal, state and local law enforcement officers.

Preaskorn said NGI improved the accuracy of fingerprint algorithms to 99.6 percent from 92 percent. Officers now have access to the Repository for Individuals of Special Concern (RISC) through their mobile devices.

“An officer can take one or two fingers and get images back to us and within six seconds or so, we will give a probable cause, red light or green light type of thing,” he said. “We redid our latent and became three times more accurate. We went from about 26 percent accurate on our latent submissions to about 82 percent to 86 percent accurate.”

Preaskorn said the IAFIS system didn’t process flat or latent fingerprints as opposed to rolled fingerprints well.

Most recently under increment four, the FBI received facial recognition capabilities and a Rap Back services, which provides law enforcement officers notification of criminal, and, in limited cases, civil activity of individuals that occurs after the initial processing and retention of criminal or civil transactions.

“We take still images, which are mug shots and photos contributed by law enforcement, and we will return a candidate list and they will have a gallery to select from,” Preaskorn said. “We also are scalable now. We built this on a service-oriented architecture so as a modality changes we can unplug in and plug in a new vendor, or expand if we want to add iris in the future. We can just plug it into the service-bus and it allows us to expand and grow.”

Probably the biggest difference between the two systems is the FBI designed the IAFIS system for about 62,000 transactions a day. The bureau created NGI to handle as many as 700,000 transactions a day.

“We’ve actually hit 693,000 on April 16. It’s not a typical day. The busy times are 260,000 or so,” he said. “It’s scalable, more flexible built on blade technology, increase core processing, add memory or change hardware at will.”

Other feds using NGI

Preaskorn said by building NGI increments over the last few years, the FBI allowed field agents and other stakeholder to test and comment on the system. This ensured the FBI wouldn’t suffer from scope creep and that it was meeting users’ needs.

The FBI will continue evolving NGI over the next few years. Preaskorn said they are piloting iris technology to see if it’s viable. He said the bureau also is considering DNA, voice or other modalities as those technologies improve.

The Next Generation Identification System isn’t just for the FBI. Preaskorn said the Homeland Security Department is its biggest federal customer. DHS sends prints to NGI from all border crossings and port of entry checkpoints.

The Defense Department and the Office of Personnel Management also rely on the FBI to process fingerprints against NGI.

“I think it’s a huge recognition for a lot of the hard work and effort that the whole team put together,” Preaskorn said. “Not just myself, my fellow employees, my partners, my peers, Lockheed Martin and all the subcontractors, a lot of hard work has gone into this and it’s really nice to be recognized by an agency like this.”

RELATED STORIES:

Lockheed lands FBI biometrics contract

FBI deploys 30,000 Android smartphones to field offices

FBI ready to give Next Gen ID a whorl

Copyright © 2019 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.