A high-resolution photograph of an iris, the colored part of the eye, contains 240 unique points of reference for identifying a person — that’s about 100 more reference points than a fingerprint, according to the FBI. Although iris recognition is nowhere near replacing fingerprint technologies and databases, it’s become a popular way for law enforcement and corrections agencies to add another layer of certainty when identifying offenders.

More than 2,100 departments in 27 states use iris recognition, according to media reports from 2007. And in November 2010, the technology received high-profile coverage when the New York City Police Department (NYPD) began using it. Paul Browne, the NYPD’s chief spokesman, said via e-mail that two incidents drove the department’s use of the technology: “In early 2010, where [two] individuals awaiting arraignment escaped incarceration by feigning to be individuals being arraigned for lesser offenses,” he wrote.

One man was charged with five robberies but posed as a prisoner arrested for a small-time drug offense; he was released on his own recognizance by a judge.

As citizen safety is a prime concern for law enforcement agencies, iris recognition has reduced the possibility that the wrong person is released from jail, which can be devastating to a community.

Positive Identification

According to a National Science and Technology Council document, the color and structure of an iris is genetically linked, but the pattern isn’t. “Prior to birth, degeneration occurs resulting in the pupil opening and the random, unique patterns of the iris,” the document says. Iris recognition systems analyze this random pattern.

Iris imaging relies on high-quality digital cameras that typically use infrared light to illuminate the iris, a method that is supposed to be painless for the individual. Many law enforcement agencies that have adopted the technology photograph the iris during the booking process and have handheld devices that allow police officers to verify an individual’s identity at any location or time, for instance, when an offender is being released from jail or transported to court.

The NYPD’s $500,000 program was funded through the U.S. Department of Homeland Security, and was rolled out to each of the city’s five boroughs by early December 2010. Browne said the department chose iris recognition because of its speed and ease of application.

Although the NYPD’s adoption of iris recognition is fairly recent, the technology has been in use by other agencies for many years. For example, the Tarrant County, Texas, Sheriff’s Office has been using the technology since May 21, 2004. Terry Grisham, the office’s spokesman, said when the technology went live nearly seven years ago, officers took images of the irises of each person in the jail at that time, which was between 3,000 and 4,000 people. “We went throughout the jail, iris scanning everybody in custody,” he said. “And then from that point on, when someone came into jail, they were iris scanned and it went into the database.” The jail’s population averages around 3,500 people per day, and today the office has almost 230,000 unique iris scans in its system.

Now when someone is being booked in the Tarrant County Jail, iris scanning is part of the process. A handheld system scans each eye, and the visible characteristics are converted into a 512-byte Iris Code, which is represented as a coordinate system that looks like a series of bar codes. Because there isn’t a national database for state and local law enforcement agencies to verify the iris data when identifying people, the jail must rely on its own stored scans. Grisham said the jail has numerous “frequent fliers” — approximately 50 percent of the jail’s inmates are repeat offenders. As other agencies enlist the technology, the county intends to develop a system so officers can share information.

Before adding iris recognition to its booking system, the process for the Tarrant County Sheriff’s Office to positively identify an offender was arduous. The office electronically sent offender’s fingerprints to Austin, Texas, for comparison in the FBI’s Automated Fingerprint Identification System. On a good day, Grisham said, it takes as long as four hours to get a return on that information. “Because we don’t have the fingerprint equipment in the housing units, the inmate has to be brought down to the release area and fingerprinted,” he said. “Then in the old days, we would have to wait — that inmate would sit for as long as it would take us to get a positive return back because we don’t just rely on local databases, we want everything.”

Iris recognition doesn’t replace fingerprint scanning, but it adds another layer to the Sheriff Office’s identification matrix. Other identifiers the office uses include photographs and information about unique markers, like scars, tattoos and missing body parts. “They all kind of go together to make a positive ID,” Grisham said.

The technology has reduced the possibility that the wrong person will be released from jail, he said, adding that in the last 10 years, fewer than five offenders have been released by mistake.

Eyes on the Future

As using iris recognition to identify offenders becomes more popular, standards for its use become increasingly valuable. The National Institute of Standards and Technology (NIST) is working toward the standardization of iris data, said Patrick Grother, a NIST computer scientist, so that a marketplace can be built around it. Standards, he said, would make it feasible to place an iris image on a passport. Then an individual with a passport could travel to another place in the world, and officials there could read and understand the iris image, potentially removing any chance of misidentification. “We’re also quite active in testing to make sure that it works,” Grother said. “We engage the iris recognition community, commercial and academic, to test algorithms and to test the ability to recognize things. And like all biometrics, sometimes things go wrong; no biometric is perfect.”

In January, a new standard was developed, ISO/IEC19794-6, that specifies data interchange formats, which Grother described as the “MP3 file of iris recognition.” Another standard that’s under development, ISO/IEC 29794-6, will appear in 2012 or 2013 and will regulate the properties of an iris image, including its focus, area, exposure and contrast. “All those properties that make a good iris a good iris,” said Grother.

Iris recognition is a rapidly evolving field, but still won’t be a perfect identifier for all offenders. Grisham said sometimes the Tarrant County Jail’s booking technicians can only scan the iris of one eye due to someone’s ocular bone structure, and when people are under the effects of drugs or alcohol it can be difficult for them to open their eyes wide enough or to stand steady while the images are taken.

For agencies interested in using the technology, Grisham said to convey the importance to funding organizations that the release of the wrong person from jail can be devastating to a community. “If it saves one person from being out on the streets who shouldn’t be there,” he said, “then in our minds, it’s worth the investment.”

Elaine Pittman  |  Associate Editor

Elaine Pittman is the associate editor for Government Technology, Public CIO and Emergency Management. Before coming to Government Technology, she worked for The Coloradoan daily newspaper in Fort Collins, Colo. She can be reached via email and @elainerpittman on Twitter.