The FBI was certain its new fingerprint processing system would help solve cases faster.
But a year into using the Next Generation Identification system (NGI), the bureau is reaping benefits beyond what it imagined five years ago when it first began the update of its key crime solving service.
Stephen Morris, the assistant director of the FBI’s Criminal Justice Information Services (CJIS) Division, said from “soup-to-nuts” NGI has led to faster and more accurate processing of fingerprints, including a rise in the reopening of cold cases.
“From the deployment of some of the new technology, the larger screens and faster processors, we realized the efficiencies from day one. As we rolled that equipment out, it made it easier for examiners to process the current workload that the legacy Integrated Automated Fingerprint Identification System (IAFIS) was generating,” Morris said. “As we rolled out every increment — and that’s kind of the beauty of an incremental, agile development — as you deliver increments you realize the efficiencies from those before you move to the next increment, so it’s kind of like a building block. So with every increment we deployed, we had success and efficiencies that we gained.”
Now that all five increments have been in place for more than a year, the FBI and its federal, state and local law enforcement partners are experiencing significant improvements.
For example, the FBI says NGI helps law enforcement officials process 650,000 criminal transactions daily versus just 160,000 a decade ago. The response time for searches today generate responses within 10 seconds, down from 2 hours in 2000.
And NGI’s accuracy rate is on average 99.6 percent, up from 92 percent under IAFIS and the older technologies.
“When we deployed the search algorithm during a five-day period back in 2011, we ran the old and new systems in parallel. All transactions went into both systems and the new search identified about 910 candidates that the old one missed,” Morris said. “That is a huge improvement. It was the beginning of a number of successes that exceeded our expectations. We also improved our latent search algorithm. Latent prints are harder to process than standard flat fingerprints. The accuracy and results with the new latent algorithm, which we started using in the middle of 2013, was three times more accurate.”
More than just a fingerprint system
NGI is more than just a fingerprint system. It also provides law enforcement officers with access to palm prints, latent fingerprints including those gathered at a crime scene, and facial recognition. The FBI currently is testing iris recognition technology as well.
“There are many examples out there where we kept law enforcement officers safe by identifying these individuals with criminal records or are wanted for various crimes, so when you think about those law enforcement officers approaching a vehicle, they are much more aware today with NGI than they’ve ever been in the past of what might be waiting for them out there,” said Jay Winkeler, Lockheed Martin’s director of justice technology solutions and services. “So getting those law enforcement officers the information they need to stay safer and by definition, they are able to take those bad guys off the street and keep the community safer as well.”
The FBI hired Lockheed Martin as the prime contractor in 2008 under a 10-year, $1.1 billion deal. Lockheed delivered the program on-time and on-budget as of September 2014.
Winkeler said all of these modalities and the tools associated with NGI are enabled by a flexible and scalable technology architecture.
He said the agile approach enabled the FBI and Lockheed to collaboratively make sure NGI was progressing as expected.
“We delivered capability out to the user community and gauged how that capability was fielded and if it was meeting the needs of the organization, and allowed us to go back iteratively to tweak things along the way and make sure we were delivering the right capability to the users at the right time,” Winkeler said. “We really took full advantage of that agile methodology there, such that in the old system it would take four or five years sometimes before you delivered the first iteration of the system.”
The future of NGI
Morris said the future of NGI will include mobility and multi-modal technologies, such as iris or voice.
“We are piloting iris technology today and the accuracy is relatively high, but some of things we are seeing is iris is more applicable in a static environment where you have more control over the individual you are searching, like a prison setting or courtroom or other areas where you have control of flow of people,” he said. “In prisons and different organizations involved in corrections, there is a tremendous amount of potential.”
The FBI also has been testing a mobile fingerprint scanning capability of one or two fingers using something that looked like a bulk cell phone, Morris said.
“It would run a search of what we refer to as Repository for Individuals of Special Concern (RISC) where we took the worst of the worst out of National Crime Information Center (NCIC), people who are wanted for murder, or have arrest warrants or are sex offenders or violent offenders,” he said. “So we reduced the holdings the system had to search and we provided a successful search experience with mobile devices. We know it’s impractical to search 100 million prints, so we distilled the database down to worst of the worst and we could return responses to an officer in less than a minute.”
Morris said now officers can connect NGI through these mobile devices and are performing about 2,000 searches a day that take on average about four seconds.
The FBI also makes NGI available to other agencies, including the Homeland Security and Defense departments.
“We are constantly emphasizing when we talk to private industry or vendors about developing new technologies, the key to the future in all of this biometric technology is interoperability,” he said. “That is something we have been involved in for well over 10 years and because of the success of NGI, like anything else, more people want to use it so interoperability is even more important.”
Winkeler said Lockheed continues to work with the FBI to improve NGI on the back-end and the front-end of the system.
He said both facial recognition and iris matching algorithms are two ways to provide a more comprehensive match of identity.