U.K.'s biometric trial exposes 'teething problems'

Fat fingers sometimes stumped the equipment

As part of its campaign to relaunch legislation for a national identity card program using biometric technology (see story), the U.K. government late yesterday released its report on the U.K. Passport Service's (UKPS) eight-month trial of biometric technology involving 10,000 volunteers.

Rather than building a solid case for the Identity Cards Bill, the report highlighted what the UKPS called "teething problems," including complications with scanning fat fingers, dark-complexioned people and older individuals.

The UKPS stressed that the trial's aim was to measure people's reaction to having biometrics data collected, not the technology's effectiveness at gathering facial, iris and fingerprint information. Nevertheless, the results revealed some glitches.

The U.K. has committed itself to begin including facial image biometric identifiers in new passports starting next year. The government now intends to require a second-generation biometric passport using fingerprint identifiers beginning in 2008, partly because the facial biometric will not keep people from registering more than once.

While 10,000 volunteers participated in the trial, the detailed research released by the UKPS focused on 2,000 "quota" people picked to match the general population, and 750 disabled people.

Of the three biometrics tested, the lowest verification success rate occurred with the face, with a 69% success rate for quota participants and a dismal 48% for disabled participants. However, the UKPS attributed that rate to the lighting in the mobile enrollment center, rather than the fact that the people involved in the trial have a disability.

The verification success for fingerprints was higher, at 81% for quota participants and 80% for disabled participants. The UKPS pointed out that the fingerprint devices used in the study increased the failure rate because they occasionally failed to record sufficient details. The machines also had difficulty scanning the irises of dark-complexioned people and those over the age of 59, for which the report recommended further studies.

Fat fingers appeared to pose occasional problems for the fingerprint scanner, as the device was sometimes "too small to scan a sufficient area of fingerprint from participants with large fingers." In addition, fingers with scarring reduced the rate of verification success as well, the report said.

Not only did the iris biometric have the highest success rate among the three tests, at 96% for quota participants and 91% for disabled participants, it also proved to be the preferred biometric for both males and females. However, many of the females tested also preferred fingerprints, the UKPS said.

Atos Origin SA ran the trial and delivered and installed the project's equipment and software. NEC Corp. supplied its Automated Fingerprint Identification System; Identix Inc. handled the fingerprint capture and facial matching technology; and Iridian Technologies Inc. provided the iris recognition technology.

The government has not awarded the biometric passports contract, and a spokeswoman said she knew of no timeline for doing so. Atos declined to comment.

The UKPS report did not look at the reasons behind the findings or offer any suggestions for technology fixes. "These may be addressed in further trials," the report said. But there are currently no plans for further biometric trials.

Another potential headache for the UKPS could be the amount of time it takes to physically gather the biometric information from individuals. According to the report, it took an average of 7 minutes, 56 seconds for the quota group to be enrolled and an average of 8 minutes and 15 seconds for the disabled group.

There are currently about 60 million people living in the U.K., with approximately 80% of the adult population currently holding a passport. Between 3 million and 4 million people renew their passports each year, the UKPS said. Passports can be renewed through mail, but once biometrics are used, individuals will be required to show up at centers that will collect biometric information.

The U.K. government has always planned to use the UKPS effort to build the base for the ID card plan and its resulting database. The aim is to issue the ID cards in 2008 in conjunction with passports at an estimated cost of $170 for the pair. The price of renewing a 10-year passport in the U.K. is now less than half that amount.

The Liberal Democrats, the smaller of the two U.K. opposition parties, pointed out today that because the estimated cost for the combined 10-year ID card/passport package omits start-up costs and cross-subsidies of free ID cards for pensioners and the poor, the final price per unit could end up being higher.

The government wouldn't comment other than to say that its current estimates are only for how much it costs the UKPS to produce the ID card/passport unit, not necessarily how much an individual would be charged.

Copyright © 2005 IDG Communications, Inc.

Shop Tech Products at Amazon