• silicon.com
• Technology
• Security

By Peter Cochrane, 27 March 2006 10:30

COMMENT

Written in Washington Dulles Airport after a heavy day, polished on the M5 between Reading and Taunton some weeks later, and dispatched from a domestic wi-fi link in Colyton, Devon

As a regular international commuter across numerous countries I often pass through security points of varied intensity. Some are extremely lightweight, others extremely thorough, and most just plain procedural with a tendency to be more concerned about process than security itself.

In the US I regularly have to give two finger imprints and have my face recorded on entry to the country. The system seems quite slick, it works well, has been implemented at a fairly minimal hardware cost and the software seems to do the job. The agents always have my details before them including my entry and departure record in and out of the US. Armed with the latest information they can ask any pertinent questions. However, what this system has not done is speed up my entry, in fact it has actually slowed it down.

There have also been various experiments in the US with hand and thumb print readers so that people can pass quickly through security on the basis of a continual vetting of their security status but only for internal flights.

You would think the international airlines, and nations across the world, would recognise that a traveller with 35 years of trouble-free flying, working for internationally recognised companies and corporations would be a sure-fire bet for immediate entry at high speed. But no, the same amount of security energy is applied to everyone. In fact, in some countries there is an apparent bias away from those who may most likely be terrorists, towards people who are most likely not going to be on the basis of some politically correct notion of avoiding various forms of harassment.

So what's the answer? The UK government has just delayed the introduction of iris scanning at all UK airports for reasons concerned with the technology rollout. But I look forward to the arrival of this technology (soon please?) on the basis that it should allow me to walk straight through at the blink of a camera.

More than 10 years ago I was involved in the development of iris scanning and was very impressed by the fact that it was about six orders of magnitude more accurate that taking a human DNA sample.

So what's the big deal? All you require is a camera, a PC and the software. It can't be fooled by photographs because you can take into account curvature of the eye. It can't be fooled even by a dead eye or a severed head because the annular muscle around the iris has to be seen to be pulsating. The iris pattern is set from about three months of age and remains so for the rest of life, barring the incidence of ocular deformations and damage, which can also be accommodated by other techniques.

Will people object to this system? Possibly - but it can get very sophisticated. For example, it is possible to scan the human iris on the move without people knowing that it has been done in the same way that you take photographs of people with your pocket camera. And it seems to me that covertness of security is rather more assuring than the overt systems that demand I take off my trouser belt, my shoes, empty my pockets, remove my coat, take my laptop out of my bag - time after time after time...

Hopefully we will see a compromise across the world with security systems that incur a minimal inconvenience at a maximum probability of catching miscreants. There are many techniques we can use that including fingerprints, or better still (perhaps?) the vein structure in the thumb and hands. Just like the iris, these structures are unique to each person. You can try using facial and/or voice recognition but by comparison these are very crude and relatively unreliable.

However, if you concatenate a series of very low cost biometrics that each gives you a poor error rate, it is possible to achieve a very high performance. For example: voice plus face plus finger recognition give a combined error rate around 10exp(-18), which is considerably less that a good iris scanner at around 10exp(-24) but it is also an awful lot cheaper!

For my part, and for the sake of all other travellers worldwide, I just wish a global standard would get agreed so we can all get back to travelling instead of undressing!

Personally, I'd go for a near 100 per cent covert solution with a minimal inconvenience for those who just want to travel and do business, and a maximum opportunity to intercept those deranged beings intent on doing evil. For now, and possibly forever, I'm really happy to waiver all my rights to privacy provided we are all safe as a result.