The so-called super-recognizers They are exceptional at memorizing faces. But what is your secret?
A new study, conducted by researchers at the University of New South Wales (UNSW), in Australia, reveals that people who never forget a face work in a “smarterno more intense.”
In other words, they naturally concentrate on more distinctive facial features of each person.
Is this good news for the rest of us, who would like to avoid future embarrassments caused by confusing someone we already know? Unfortunately, not really.
“Your aptitude is not something you can learn like a trick”, explains the psychologist James Dunnresearcher at UNSW and lead author of , which was published in Proceedings of the Royal Society B: Biological Sciences. “It is an automatic and dynamic way of capturing what makes each face unique”.
To understand what people see super-recognizersDunn and his team turned to technology eye tracking to reconstruct the way different people observed unfamiliar faces.
The study involved 37 super-recognizers and 68 participants with common facial recognition capabilities, analyzing where and for how long each one stared at photographs displayed on a screen.
The study authors then fed this data into machine learning algorithms trained to recognize faces. These algorithms — deep neural networks — were tasked with determining whether two images whether or not they belonged to the same person.
“AI has become highly competent at facial recognition”, says Dunn, quoted by . “Our goal was to take advantage of this to understand which human gaze patterns provide the most information.”
It is clear that our brain plays a decisive role in processing visual information. But when the algorithms received eye-tracking data from the super-recognizers, got it right more often in matching faces than when working with data from participants with normal abilities.
“These results suggest that the perceptual underpinnings of individual differences in the ability to recognize faces may originate in the earliest stages of visual processing — at the level of retinal coding,” Dunn and colleagues write in the paper.
The study expands on previous research by the same team, which showed that super-recognizers transform a face into a kind of puzzle: they divide new faces into parts, before the brain puts them back together into a composite image.
This puzzle-like approach goes against the idea that memorizing faces effectively depends mainly on look at the center of the face and see it as a whole.
The new research expands on these findings, suggesting that super-recognizers don’t just gather more information than the average mortal — they focus especially in the elements that carry more “clues”.
“It’s like in the caricatures: when the most distinctive features of a face are exaggerated, it becomes easier to recognize it“, explains Dunn. “The super-recognizers seem to do this visually, they tune in to the most revealing traits of a person’s identity.”
This line of research could help improve facial recognition systems, although the authors emphasize that, for now, humans still have an advantage about AI, because they draw on other cues present in social interactions.
However, we should not assume that we are exceptional. There is robust evidence that there is a strong genetic basis for the ability to memorize faces with great precision, but the processing of facial identity also plays an essential role in primate social behavior, so the biological roots of this ability may not be exclusively human.
