‘Unwarping’
As biometric identification is rolled out in everything from passports to credit cards, recognising fingerprints is no longer just for police departments. But the problem, as we know from TV detectives, is that often you get distorted prints. Now scientists at the University of Warwick have developed a technology, called ‘unwarping’ that allows even distorted fingerprints to be identified.

Biometric identification normally requires a user to register in a database, and when the user pushes their finger onto a surface for identification, the system goes through all the fingerprints in the database to see whether it matches. But this can be cumbersome, slow and inaccurate.

Fingerprint variations
Dr Li Wang, a researcher in the University of Warwick's department of computer science, explains that unlike other forms of ID recognition, like keying a PIN number on your credit card, ‘the difficulty for a biometric machine is the variability of the data. Every time you present your finger to a scanner, you do it in a slightly different way which can be due to an environmental factor, or unintentional variation pressures, temperature, skin moisture, lighting, these can all be very different every time.’

The method developed by Dr Wang and his colleagues at Warwick means that not only can distorted and incomplete fingerprints be identified, but it’s done almost instantaneously. ‘We had to come up with a way to automatically detect those distortions and subsequently remove them. We call that process normalization, we create a ‘geometric invariant’. It doesn't matter how much pressure is applied, or at what angle, or in what position, we can correct that and a make a unique representation of the fingerprint.’  

Science museum test
The team has already carried out successful public tests. ‘We have collected fingerprints for our internal database,’ explains Wang. ‘We've tested on the campus and at the Science Museum in London. We wanted to test it in an uncontrolled environment and it worked brilliantly.’

Because it’s done at such speed, this technology is incredibly valuable, whether presenting a passport or using fingerprints to enter a building. ‘It could be used in mobile devices, for personal usage,’ says Wang. ‘Or it could be in a very large scale scheme such as ID cards. We believe the technology has to roll out to all of them.’