Unlike its original electronic fingerprint database that went more than a decade without upgrading its fingerprint matching software, the FBI’s new multimodal biometric database is being enhanced along the way, taking advantage of the flexibility built into system, an agency official says.
The FBI’s original Integrated Automated Fingerprint Identification System (IAFIS), launched in 1999, didn’t receive any upgrades to its 10-print and latent print algorithms until the first increment of the replacement system was deployed in 2011, according to Kim Del Greco, deputy assistant director for the agency’s Criminal Justice Information Services (CJIS) Division. She says the 10-print and latent algorithms that IAFIS launched with were based on designs from 1992 and 1993 respectively.
When the new biometric identification services system—called Next Generation Identification (NGI)—achieved initial operating capability in 2011, the system was built so that upgrades could be introduced relatively quickly and additional modalities could be added.
“NGI was designed to be flexible and to add and delete enhanced biometric modalities,” Del Greco says at AFCEA’s annual Federal Identity Forum. “It was also designed to change algorithms without a costly reengineering effort.”
Initial operations began with the upgraded fingerprint matching technology and by the time full operational capability was achieved in 2014, NGI featured a repository for individuals of special concern, latent fingerprint and palm print capabilities, face recognition services, and a Rap Back system that allows for authorized users to receive updated status notifications on any criminal history reported on certain individuals like school teachers.
Since full operational capability, NGI has already received key upgrades.
Del Greco says that last fall enhanced 10-print capabilities were introduced providing more accuracy, flexibility, scalability and reliability. At the same time, open source software was added into the baseline database, she said.
A spokesman for the CJIS division tells HSR that NGI has always had open source products but the agency “is continuously looking for opportunities to maintain system performance while reducing system operating costs by incorporating additional open source products.”
This fall the FBI is upgrading the latent print capabilities that will offer similar improvements that the fingerprint enhancement provided, Del Greco says. The improvements to the system augur new opportunities to support and close cold cases around the country, she says, leading the agency to begin meeting with law enforcement agencies and courts nationwide to resubmit some of these cases.
The NGI system was developed a business of Leidos [LDOS] that the company acquired from Lockheed Martin [LMT] more than a year ago. Leidos still supports NGI. The biometric matching algorithms are supplied by OT-Morpho through its U.S.-based entity, Trans Digital Technologies. OT-Morpho is owned by the private equity firm Advent International.
In 2018, the FBI plans further enhancements to the 10-print and latent print algorithms and will make 10-print segmentation upgrades, Del Greco says. The agency is also looking to make improvements to NGI’s facial recognition capabilities next year, she says. NGI contains 6.1 million face records.
“We don’t want to rest on our laurels,” Del Greco says. “We want to invest each and every year in the latent, the 10-print and the facial algorithms.” These investments take the form of either trade studies or upgrades and enhancements, she adds.
The FBI also has a four-year old iris recognition pilot and is developing standards for capture, matching and transmission, Del Greco says. The business case is being developed this fall to help with onboarding iris capabilities for larger scale operations, which are expected in the next two years, she said. The FBI is also focusing on certification of iris devices, she adds.