Some People Are Wired to Spot Faces Instantly — And Their Skill Is Shaping Smarter Tech
Super recognizers are people with an almost supernatural skill to quickly identify thousands of individuals based on their facial features, an ability that has also inspired law enforcement to fill their ranks with these rare few.
To understand exactly what super recognizers do differently, researchers from the School of Psychology at the University of New South Wales (UNSW) tracked their eye movements while analyzing facial images and trained AI with those patterns. The results, published in Proceedings of the Royal Society B: Biological Sciences, were surprising.
“Super-recognizers don’t just look harder, they look smarter. They choose the most useful parts of a face to take in,” said study lead James D. Dunn in a press statement, who hopes his team’s research sheds light on human visual expertise and helps refine facial recognition technology.
Read More: Airports Are Embracing Facial Recognition. Should We Be Worried?
What Makes a Super Recognizer
Some people among us are exceptionally good at recognizing faces, giving them the ability to identify strangers they’ve only seen once — even if their appearance has changed or they’re captured on a low-resolution CCTV camera.
This superior face recognition ability has fascinated researchers for years, who continue to explore its origins and prevalence in the population. They believe it’s likely genetic and varies across individuals. To be considered a super recognizer, a person must score exceptionally high on multiple facial recognition tests developed by experts in the field.
By studying how super recognizers collect visual information using eye-tracking technology, scientists found that it’s not just about the quantity of information. Instead, these individuals seem to have an innate ability to pick up on the right clues from someone’s face, allowing them to build the clearest mental picture possible.
Remembering Key Facial Features Matters Most
To test their theory, the UNSW team compared the performance of 37 super recognizers with 68 people who had average facial recognition skills, using eye-tracking technology to monitor how they observed faces on a computer screen. The collected data was then shared with AI networks trained to identify faces.
“AI has become highly adept at face recognition. Our goal was to exploit this to understand which human eye patterns were the most informative,” explained Dunn in a press statement.
The researchers then tested two differently trained AI models: one trained on data from super recognizers and the other on data from average participants. Unsurprisingly, the super recognizer-trained AI was better at matching faces. But what stood out was that even when both models received the same amount of information, the super recognizer AI was still more accurate. This suggested that remembering certain facial aspects is more important than simply memorizing as much detail as possible.
Improving Facial Recognition Technology
Unfortunately, no matter how much you train, super recognition isn’t a skill that just anyone can learn. Dunn said it’s not simply about what you look at and where, but about how the brain processes that information in an automatic, dynamic way.
“It’s like caricature — the idea that when you exaggerate the distinctive features of a face, it actually becomes easier to recognize. Super-recognizers seem to do that visually — they’re tuning in to the features that are most diagnostic about a person’s face,” he added.
Today, facial recognition systems in controlled environments, such as airports with high-quality images and stable lighting, already excel at identifying faces. But under less ideal conditions, humans with exceptional facial recognition abilities still have the edge, Dunn explained. Their research could help close that gap, inspiring new ways to improve facial recognition software.
Read More: What is Prosopagnosia Or Face Blindness?
Article Sources
Our writers at Discovermagazine.com use peer-reviewed studies and high-quality sources for our articles, and our editors review for scientific accuracy and editorial standards. Review the sources used below for this article: