Feb. 2017 Digital Edition
January 2017 Digital Edition
Nov/Dec 2016 Digital Edition
Oct 2016 Digital Edition
Sept 2016 Digital Edition
Aug 2016 Digital Edition
July 2016 Digital Edition
Iris recognition is being tested on illegal immigrants in McAllen, TX
The growing importance of “multimodal” biometrics in the security field -- in which more than one biometric method is used to capture identifying information about the same individual – is demonstrated by an important 60-day pilot program has just been launched by DHS at a U.S. Border Patrol facility in McAllen, TX.
The biometrics team at US-VISIT, which is part of DHS’s National Protection and Programs Directorate (NPPD), began on July 31 what it calls a “limited production pilot” to test the efficiency, speed and operational effectiveness of two different types of iris recognition equipment as well as two different types of facial recognition gear. Using these biometric systems, Border Patrol employees will capture images of approximately 200 alleged illegal aliens per day (or about 12,000 aliens during the course of the two-month pilot), explained Will Graves, the chief biometric engineer helping to oversee the US-VISIT project. While the results from the facial recognition portion are important, Graves told GSN at a biometrics conference in New York City on August 1, “We are focusing primarily on iris recognition.”
The goal of the iris recognition pilot is to determine the fastest and most accurate ways to capture images of an individual’s iris that could be added to the department’s growing IDENT database and could be used to identify that same person at some point in the future. Biometrics experts have become convinced that using two separate biometrics technologies, and comparing their results for a single individual, can increase the accuracy of the identification.
US-VISIT maintains two massive databases: IDENT, which stands for Automated Biometric Identification System, and currently contains 140 million unique individuals in its system, said Graves, and ADIS, which is short for Arrival and Departure Information System, and is comprised principally of flight manifests of arriving and departing foreigners, which enable DHS analysts to determine which of those travelers has “over-stayed” his or her authorized time in the United States. ADIS contains records of 250 million individuals, but these do not include any biometric images, Graves explained at the conference, which was organized by the Advanced Learning Institute, Inc., of Oak Brook, IL.
At the McAllen, TX, pilot, DHS researchers are testing three different types of iris recognition equipment: handheld (which can be aimed at the individual’s iris by border patrol personnel from a very short distance); “on-the-move” (which try to locate the iris and capture a workable image while the illegal alien is still a long way from the counter; and “stop-and-go” (which require the person to stop, and perhaps even change directions, in close proximity to the counter.)
During an earlier version of the same pilot program in McAllen, approximately 500 people were imaged during a two-week period. DHS learned several important lessons during that earlier pilot, Graves reported. Perhaps the most important lesson was that tried-and-true mantra, KISS -- Keep It Simple, Stupid. Graves recalled that border patrol employees and the illegal immigrants they imaged all performed most effectively when the biometric devices and instructions were kept as simple as “tinker toys.”
For example, getting the immigrants to stand the proper distance from the camera was easy once the researchers taped the outline of feet on the floor. “If you put feet on the ground,” recalled Graves, “they will stand there like lemmings.” Training the border patrol personnel to aim the handheld cameras correctly was also a piece of cake. “Habituation was not a problem with the device,” Graves added.
One of the key findings from the earlier pilot: “There are a lot of abnormal eyeballs in the real world,” observed Graves. While small-scale tests in a lab might not come cross such abnormalities, larger-scale efforts using real-world individuals can turn up some surprises. For example, Graves and his colleagues noted a surprising number of darkened streaks appearing on people’s irises, which sometimes made it impossible to capture a usable image. “In the third world, where there is a lot of sunlight, there can be a lot of these streaks,” he explained.
Ultimately, when the latest McAllen pilot is completed and the iris recognition results are analyzed, the question will merge: Does any U.S. Government department, agency or office plan to deploy iris recognition systems on a large scale to accomplish their mission? At this point, there is no clear answer to that question.
Many agencies might acknowledge the ever-increasing sophistication of iris recognition technology, but not possess the budget or the political wherewithal to try to implement such a cutting-edge program.
Perhaps the most logical adopter will be the U.S. Border Patrol itself, which has been participating right along in the pilot programs in McAllen.
The Border Patrol is interested for several reasons, says Graves.
First, such a stand-off biometrics technology would require border patrol agents to have less hands-on contact with the illegal immigrants being detained. “They like iris because they don’t have to touch the individuals to capture their biometrics,” Graves said.
Second, the collection of iris images could help create a local watchlist of illegal immigrants, which could facilitate the identification of repeat offenders, if they show up along the border again and again.
Third, iris images can strengthen the ability of the Border Patrol to maintain an incontrovertible “chain of custody” that could distinguish those detainees who are being held by U.S. authorities and those detainees who have been released.