Photo-ID security checks flawed: Study

Dr. Rob Jenkins from the Department of Psychology at the University of York said: “This level of human error in Australian passport office staff really is quite striking, and it would be reasonable to expect a similar level of performance at U.K. passport control.

“At Heathrow Airport alone, millions of people attempt to enter the U.K. every year. At this scale, an error rate of 15 percent would correspond to the admittance of several thousand travelers bearing fake passports.”

U.K. passports are valid for a 10-year period and as a result officers also have to take into account changes in a person’s appearance over time.

In a second test, the passport officers were asked to match current face photos to images taken two years ago or to genuine photo-ID documents including passports and driving licenses.

Error rates on this task rose to 20 percent — a level of performance that was no different to a group of untrained student volunteers who were also tested.

Dr. David White of the University of New South Wales in Australia, lead author on the paper, said: “While it might have been expected that years of training and experience would have improved passport officer performance, our study showed this was not the case. Passport officers were no more accurate than university students.”

Professor Burton added: “This study has importantly highlighted that the ability to be good at matching a face to an image is not necessarily something that can be trained.

“It seems that it is a fundamental brain process and that some people are simple more adept at it than others.

“Our conclusion would be that focusing on training security staff may be ploughing efforts in the wrong direction. Instead we should be looking at the selection process and potentially employing tests such as the ones we conducted in the study to help us recruit people who are innately better at this process. Because of this study, the Australian Passport Office now set face matching tests when recruiting staff and when selecting facial comparison experts.”

The study of Australian passport office staff adds to research being conducted at the University of Aberdeen asking if security measures would be enhanced if passports carried more than one image of a person.

The researchers say they hope the research — which involves collaboration with worldwide passport controls — could lead to changes in how security systems operate in the future.

Professor Burton continued: “This separate study is examining if having a multitude of images taken under different conditions presented on a passport would increase precision in facial recognition.

“What has been missing in the development of security technology so far is the fact that one photograph does not give us a true representation.

“There is a great emphasis on a passport image to fit all purposes but people often comment on the fact that their passport photo looks nothing like them? This observation turns out to be true when tested scientifically.

“Findings from our studies show that what really matters when you learn to recognize someone is the range of pictures you see — all the possible ways a person can look in photos.

“It seems strange that we expect a single passport shot to encompass a person and allow us to consistently recognize them. If we are stuck on the concept that a good representation of a person is achieved through one image, then we are setting ourselves up for errors.

“Could there in fact be an argument for our passports to contain a multitude of images, taken at different angles, in different lighting and formats? This is certainly something our study is examining.

“Unlocking the fundamental brain process that occurs when we see someone we know is the key here.

“If we can establish this then we can look at how we can apply this knowledge to create new technology, examine how recruitment of security staff could be enhanced and if there is potential for training techniques to support these staff.”

— Read more in David White et al., “Passport Officers’ Errors in Face Matching,” PLOSone (18 August 2014) (DOI: 10.1371/journal.pone.0103510)