The iris scanning technology can distinguish between the eyes of the living and the dead

In the film, many times we see the villain take the eyes of the dead to cross the iris scanner. Technology from movies must be more or less based on real life, right? So does a scanning system identify the human eye and the eye is dead?

The research team, including Mateusz Trokielewicz, Adam Czajka and Piotr Maciejewicz, compiled a large database after sweeping the eyes of both the living and the dead, to train the machine learning algorithm to find the difference between the two. this iris.

In their research report, they confirmed that the algorithm can distinguish the living iris and the dead eye with an accuracy of up to 99% . But besides doing iris security, it still has a hole for hackers to exploit.

Picture 1 of The iris scanning technology can distinguish between the eyes of the living and the dead
The algorithm can distinguish the living iris and the dead iris with up to 99% accuracy.

First of all, please understand carefully the iris scanning security technology . For a long time, ophthalmologists have found that the iris structure of each individual is different. Small details on the iris appear clearly in infrared scanning images, so the iris image is created by this wavelength of light appearing much in security systems.

But this system can be easily bypassed. Last year, hackers could unlock their Samsung phones by printing iris images on a contact lens, then putting glasses on a fake eye. The way to take the eye of a person who has lost it to scan (like in a good movie) is also a form through other iris scans. Until now, we have a way to discover that someone is using fake eyes.

The study to obtain the above results is based on a slightly . unusual database: it consists of 574 infrared images of iris from 17 deaths, images taken at different times later. when dying, from about 5 hours to 34 days.

The team also collected 256 iris images from other living people. They also used the same camera to scan the eye to avoid the fact that the machine learning algorithm included data about the differences between the two cameras.

There is a particularly noticeable difference between the eyes of the living and the eyes of the dead, that is, the eyes of the dead need a clamp to keep the eyes open to scan. The machine vision observation system could mistakenly recognize the other clip as the difference between the two types of eyes, so the research team cut the image carefully, only to let the iris appear.

Picture 2 of The iris scanning technology can distinguish between the eyes of the living and the dead
With this new technology, after a few hours, all eyes will be unable to pass the iris scanning security system.

The result of the study showed that the algorithm recognized all the iris died, and only a few cases received the wrong eye as the dead eye."No dead eyes were mistaken, only 1% of eyes were mistakenly received as dead," the team confirmed.

But there's a flaw: this 99% accuracy only applies to dead eyes for more than 16 hours."Eye samples that have just been collected after death (such as about 5 hours after death) can pass this test," the researchers said.

So inhumane hackers still have a way to get past the biometric system with a new eye to die (no need to imagine how they got this eye). At least there is a time to rest assured: after a few hours, all eyes will be unable to pass the iris scanning security system.