Study: MRI scans of knees can be used for biometric identification
An automated recognition system that scans crowds to identify the people within them using not their faces, or their eyes, but their knees, has been proposed.
Computer scientist Lior Shamir put the idea forward in a study published in the International Journal of Biometrics. It states that an MRI system could be used to scan the legs of people as they walk through an area, mapping the bone structure inside their knees before matching it to a biometric record using software developed by Shamir.
Shamir looked at knee scans from 2,686 people, and found that his software could recognize individuals with 93 percent accuracy. It would not be a replacement for traditional biometric identification like iris scans or fingerprints, nor standard ID such as a passport. Rather, it’s seen as more of an addition to those tactics, which could make both false readings and intentional deception statistically unlikely. Scanners could be placed alongside walkways in airports, and people could walk through them without stopping — their data would still be picked up.
The system would also have the benefit of avoiding the radiation that people are exposed to in x-ray machines. However, current MRI technology is too slow to be deployed in the field as proposed right now — a more rapid scanner will need to be developed.
Of course, biometric recognition is a controversial field. Facebook was grilled by the US government for automatically applying facial recognition technology to user photos, something it later dropped, and portrayals of iris recognition in science fiction are usually more dystopian than utopian (as in Minority Report, where adverts shout out customised greetings to passersby).
And while it is possible to sand down fingerprints and wear coloured contact lenses, it is somewhat less likely to imagine people swapping out their legs for new clean ones, without a record.