The National Institute of Standards and Technology is finding it increasingly easy to complete accurate identity checks via iris scans, the Commerce Department reports.
NIST examined 11 organizations and their techniques for acquiring and matching eye imagery.
The agency said the resulting Iris Exchange III report validates the speed and accuracy claims of iris biometrics systems.
NIST compared commercially available algorithms used for iris recognition to match individuals within a large database.
The previous method to conduct the process focused on verifying a person’s claimed identity, but systems studied in this report run the iris image through the database to conduct initial identification.
Nine private companies and two university labs submitted 92 algorithms combined through a competition.
The organizations were required to identify individuals from an eye image database of more than 2.2 million people.
The ability for the algorithms to pick out a needle in a haystack quickly and accurately is crucial, said Patrick Grother, a scientist in NIST’s information access division.
The agency found some of the algorithms were capable of searching the larger databases, which Grother says is important when there is no record to be found.
No software yielded a fully accurate result and NIST said that while some algorithms were capable of quickly searching larger databases, they were limited in terms of accuracy.
NIST said operators should practice better control over image collection to obtain better iris collections and, therefore, more accurate results.
The competition and report findings will be used to push biometric systems to 100 percent accuracy, Grother said.
The findings will also be useful to policy makers, considering incorporation iris recognition in the U.S., Canada, Mexico, India and Indonesia, he added.
A top Department of Homeland Security official told Congress in March that they would soon receive details on a biometrics system to reduce the risk of identity fraud by visa overstays.