It’s a plot straight out of science fiction: Bad guys dispose of an unlucky security guard, scoop out one of the guy’s (or gal’s) eyeballs, and hold it up to an iris scanner, fooling it into disarming a security system. As it turns out, post-mortem eyes can be used for biometric identification hours or even days after death, studies show. But if researchers at Warsaw University of Technology in Poland have their way, that might not be the case for much longer.
In a paper (“Presentation Attack Detection for Cadaver Irises“) published on the preprint server Arxiv.org, the team proposed a neural network that can tell the difference between living irises and dead ones with 99 percent accuracy.
“With increasing importance that biometric authentication gains in our daily lives, fears are increasingly common among users, regarding the possibility of unauthorized access to our data, identity, or assets after our demise,” the researchers wrote. “With a constantly growing market share of iris recognition, and recent research proving that iris biometrics in a post-mortem scenario can be viable, these concerns are also becoming true for iris.”
A typical iris scanner uses both visible and near-infrared light to take a high-contrast picture of a person’s iris, they explained. It converts patterns in these images into code, which a computer compares to a database.
Irises’ protected structure make them ideal for biometric identification — they don’t change over time, even after surgery. But it also makes them susceptible to spoofing.
The researchers began by compiling a database of 256 images of live irises and combining it with the Warsaw BioBase PostMortem Iris dataset, a collection of 574 iris images captured post-mortem from 17 people. They used a preprocessing algorithm to crop out unwanted artifacts, like the metal retractor used to hold open the eyelids of cadavers, and then compared the two datasets for signs of bias, such as differences in lighting or the angle at which the photographs were taken. Finally, they set about training a convolutional neural network — a type of neural network commonly used in computer vision — to categorize the images.
It performed well, misclassifying a dead iris for a live one only around 1 percent of the time. But there was a catch: that accuracy only held for irises captured less than 16 hours after death.
“This shows that while post-mortem iris images are relatively easy to identify, those obtained very shortly after a subject’s demise can pose problems for automatic solutions due to post-mortem changes not being prominent enough yet,” the team wrote.