Iris scanner learns to differentiate eyes from dead or alive

It has become cliché in several movies: plucking the eye of a dead enemy and going through a locked door using the iris scanner. Relying on MIT, fortunately, this kind of scene will fade from Hollywood.

The MIT project, led by researcher Mateusz Trokielewicz, does not allow the unlocking of iris scanners by the eyes of people who have died.

Ophthalmologists have long recognized that the intricate structure of the iris is unique to every individual.

To accomplish the feat, the MIT team collected images of various iris types from living and dead people, and then trained a machine learning algorithm to learn how to differentiate them. The result? The computer now has a 99% hit rate whether the eye is from a living or dead person.

It is worth noting the team's work to aim lifeless eyes at the computer: the eyes were caught on metal rods in front of the camera so that the machine learning system could work.

  • According to the team, the 99% accuracy is only valid for eyes of people who have died for more than 16 hours. That is: refinements are still required in the system

eyes

Tests

“Ophthalmologists have long recognized that the intricate structure of the iris is unique to every individual. The details are particularly apparent in infrared iris images, and iris images at this wavelength are widely used in various security applications, ”says the MIT Review. “No postmortem specimens were misclassified as live, with a probability of misclassifying a live specimen as being around 1% dead. Samples taken shortly after death may fail to provide postmortem changes that are apparent enough to serve as clues to detect vividness. ”

***

Do you know the Mega Curioso newsletter? Weekly, we produce exclusive content for lovers of the biggest curiosities and bizarres of this big world! Register your email and do not miss this way to keep in touch!

Iris scanner learns to differentiate eyes from dead or alive via TecMundo