Only 'Super Recognizers' Can Score 70 Percent or Higher on This Viral Face Recognition Test
Scientists at the University of New South Wales (UNSW) have published new research about super-recognizers: people who have a knack for remembering faces.
In their paper, the researchers describe the results of their UNSW Face Test, meant to identify super-recognizers. Since 2017, about 25,000 people have taken the test.
Their ultimate goal is to find the crème de la crème of super-recognizers to study the biological mechanisms that give rise to the phenomenon.
It's happened to you before: At the grocery store, perhaps, or at a networking event, you meet the gaze of a person you believe is a total stranger, and yet you feel in your gut that you know them from ... somewhere.
Welcome to the weird and wonderful world of super-recognizers, a rare subset of people who have an uncanny ability to recognize virtually every face they've ever seen, even from the most innocuous, boring exchanges. Scientists estimate that only about 2 percent of the human population is made up of super-recognizers, and even fewer people notice they possess this ability.
"It's actually really fascinating ... people don't always realize when they have a gift until someone says, 'Wait, you do this differently from me,'" James Dunn, a postdoctoral research associate at the University of New South Wales (UNSW) in Sydney, Australia, tells Popular Mechanics. "Sometimes, people have an awkward encounter and their friend might say, 'Do you know that person?'"
On November 16, Dunn and his colleagues published new data about super-recognizers in the peer-reviewed journal PLOS One. In their paper, they describe a face recognition test they developed back in 2017, called the UNSW Face Test. Since its inception, about 25,000 people have taken the quiz, with some 5 percent of people scoring a 70 percent or higher—the super-recognition range.
Dunn says that 5 percent figure is probably a bit high if we want to extend these results to the general population; this group of testers is self-selecting, after all. In fact, he admits he and his team have made the quiz harder than other contemporary face-recognition tests, like the Glasgow Face Matching Test and the Cambridge Face Memory Test.
That's because they're on the hunt for the best super-recognizers out there—the 1 percent. "We have geniuses, and then we have one-in-a-lifetime geniuses," Dunn says. "Bill Gates is maybe an ordinary genius, but then you have Stephen Hawking and Albert Einstein that are in the upper echelons of genius. That's what we're looking for in super-recognizers."
The Original Facial Recognition
It's easy to forget that, until fairly recently, facial recognition algorithms weren't very sophisticated. Historically, Dunn explains, humans have been far better at matching two faces than any computer program. In some highly specialized circumstances, there are actually people whose entire jobs are based on the premise of facial recognition.
To let you into a given country, border control agents must match your face to the photo on your passport, for instance. When you apply for a passport, the U.S. Department of State must determine you aren't using another person's photo before issuing your documentation. In both cases, facial recognition algorithms have their place, as the sheer volume of faces in those respective databases can be overwhelming, but humans ultimately make the final call.
There's just one problem: When it comes to matching two faces, some of the people in these facial recognition roles perform more poorly than the general population, Dunn says. Back in August 2014, his colleagues published a paper in PLOS One that analyzed Australian passport officers' error rate in face matching.
Not only did they find passport officers had no performance advantage over the general population, but in fact, the officers had a particularly high error rate. In 14 percent of the scenarios they tested, the officers accepted fraudulent photos 14 percent of the time.
That's not to pile onto people working in these roles; in all fairness, most of us aren't perfect at recognizing faces. If we were, perhaps we would all have photo identification on our credit cards, rather than signatures, Dunn explains.
But another study, published in Applied Cognitive Psychology in 1997, found that store clerks weren't so great at matching peoples' faces to their photos. Ultimately, the authors concluded, "the introduction of photographs on credit cards would have little effect on the detection of fraud at the point of sale."
These instances illustrate just how valuable super-recognizers could be. If scientists can find these people, perhaps governments or businesses could more strategically hire their face recognition specialists, leading to less overall fraud.
What Is Face Blindness?
As with so many other human characteristics, the ability to spot faces exists on a spectrum. Think of it like your typical bell curve—one tail shows the the extremely rare super-recognizers, most people fall somewhere in the middle, and those who really struggle to remember any faces at all fall at the other tail.
That last condition has a name: prosopagnosia, better known as face blindness. It's just about as rare as super-recognition, with about 1 in 50 people developing the neurological condition. In fact, scientists are confident that both face blindness and super-recognition arise from the same genetic mechanisms, but that's yet to be definitively proven.
But these people don't have poor vision or subpar cognitive abilities. Ironically, Chuck Close—an artist known for his hyperrealistic portraits, made up of tiny abstract mosaic pieces—suffers from prosopagnosia. It's particularly curious, because some research suggests portrait artists are better at recognizing faces than the average person.
In an October 2016 paper published in Vision Research, scientists at the University of Liège, Belgium found artists had better face recognition abilities than a group of average people, and could even recognize recently learned faces better than the control group. However, they didn't show any particular advantage in recognizing celebrities' faces, as that skill corresponds to long-term memory.
Unlocking the Secrets of Cognition
Scientists like Dunn know people recognize faces to varying degrees, with extremes on both ends of the spectrum. But no one completely understands why these phenomena occur. Do super-recognizers focus on facial features differently from everyone else, or is there genuinely novel activity occurring in their brains?
There's a well-formed hypothesis that both super-recognition and face blindness have something to do with the fusiform face area, a structure in the temporal lobe, located just above the right ear. This part of the brain is responsible for combining images of individual facial features together so you can recognize a person.
Dunn says his team's future research will narrow in on the goal of understanding why super-recognizers are so adept at recognizing faces. They plan to use eye-tracking technology to study how super-recognizers look at a face's individual features, for instance.
As for Dunn himself? Unfortunately, he says he has a pretty average affinity for recognizing faces.
Now Watch This:
You Might Also Like
