Ways of committing crimes have drastically evolved throughout the years; from cavemen using sticks and stones to modern men using devious tactics to steal, kill, and dupe other people. As the world continued to advance, people also found ingenious ways and developed sophisticated technology to help them protect their properties and well-being in general.
At present, you will see that most gadgets employ artificial intelligence software and various technologies, like biometrics, for instance. Some devices use fingerprint scanners; others scan your entire face; other devices just scan your iris.
As people got smarter in setting up safeguards, most thieves adapted and learned how to circumvent these safety measures. Speaking about theories, it is believed that iris scanners can be duped by using an eyeball plucked from another person. For those of you who have read Angels and Demons by Dan Brown, you will probably remember that the antimatter was stolen because the thief cuts out the physicist’s eyes and used it to fool the biometric scanners. That, indeed, is a pretty disturbing thought.
So, now, the most important question that should be asked is this: “Is there any way for iris scanners to confirm if a scanned eyeball belongs to a living person?” Recently, a group of researchers based in Poland have studied whether a machine learning system could tell between living eyeballs from dead ones.
What Exactly is Iris Recognition
To inform you a little about iris scanners, it uses iris recognition in its processes. Now, iris recognition is an automated process of biometric identification that utilizes mathematical pattern-recognition techniques on video images of one or both of the irises; its complex patterns are distinctive and stable, and it can be seen from a distance.
Some people might confuse iris recognition with retinal scanning. In order to point out the difference between the two, retinal scanning is an ocular-based biometric technology that uses the distinct patterns on an individual’s retina blood vessels and is frequently confused with iris recognition. Iris recognition, on the other hand, uses video camera technology with subtle near-infrared illumination to obtain imagery of the intricate and detailed structures of the iris that are visible.
How Dead Eyeballs can be Distinguished from Living Ones
Thanks to Mateusz Trokielewicz at Warsaw University of Technology in Poland and a couple of his colleagues, we may have an answer to that glaring question. This group of researchers has created a database of irises that were scanned from both living people and dead bodies and then trained a machine-learning algorithm to recognize the difference.
Trokielewicz says that their algorithm can differentiate a living iris from a dead one with an accuracy of 99%. Even with this high accuracy percentage, is it still possible to beat the detection system?
Efforts in Fool-proofing the System
The system was made fail-safe by means of an unusual database: the Warsaw BioBase PostMortem Iris dataset that contains 574 near-infrared iris images that were gathered from 17 people at different times of death; the dates of the images range from five hours to 34 days after death.
The group of researchers also gathered 256 images of live irises. In order to avoid future technical hitches, they were cautious enough to use the same iris camera that was used on the cadavers. They also cropped the images so that only the iris is shown in order to further fool-proof their results.
Even though Trokielewicz and his colleagues were able to yield very good results, there is still one drawback: the system’s accuracy does not apply to fresh irises. It can only produce a valid result to irises that have been dead for 16 hours or more.
Certainly, this will give thieves and criminals a window of opportunity to perform their wicked deeds, but it will also undeniably give people some comfort in knowing that plucked eyeballs will eventually lose their ability to dupe the system in just a few hours.