Super recognizers excel by looking smarter not harder

What is it that makes a super recognizer – someone with extraordinary face recognition abilities – better at remembering faces than the rest of us?

According to new research carried out by cognitive scientists at UNSW Sydney, it's not how much of a face they can take in – it comes down to the quality of the information their eyes focus on.

"Super-recognizers don't just look harder, they look smarter. They choose the most useful parts of a face to take in," says Dr. James Dunn, lead author on the research that published today in the journal Proceedings of the Royal Society B: Biological Sciences.

"They're not actually seeing more, instead, their eyes naturally look at the parts of a face that carry the best clues for telling one person from another."

Electronic eye

To find out what it is that super recognizers do differently when looking at a face, the researchers used eye tracking technology to measure where and for how long 37 super recognizers looked when examining photos of faces on a computer screen, and how that compared to 68 people with average facial recognition abilities.

With the tracking software, they then recreated what people in both groups had looked at, and fed the information into nine different neural networks already trained to recognize faces. These AI networks were then given the same task as the human participants - to decide whether two faces belonged to the same person.

AI has become highly adept at face recognition - Our goal was to exploit this to understand which human eye patterns were the most informative."

Dr. James Dunn, lead author on the research

When the researchers compared the performance of the AI in matching faces based on super recognizers' eye tracking patterns and that of average recognizers, they found a clear difference. Even when the total amount of information was the same, AI fed with super-recogniser data was more accurate at matching faces than AI fed with average recognizers data.

"Our previous research shows super-recognizers make more fixations and explore faces more broadly. Even when you control for the fact that they've looked at more parts of the face, it turns out what they are looking at is also more valuable for identifying people."

Not just a party trick

So can people with average face recognition abilities learn from super recognizers to never forget a face? Sadly no, says Dr Dunn, there's something else going on in the brain in processing the information – it's not just about where and what to look at.

"Their skill isn't something you can learn like a trick," says Dr. Dunn. "It's an automatic, dynamic way of picking up what makes each face unique.

"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."

Humans vs machines

When AI is used in the real world for facial recognition – for example, the eGates system at the airport – its processors look at us digitally and examine every pixel simultaneously, rather than looking at only parts of the face like humans do.

"In very controlled situations like eGates at the airport, where you've got stable lighting, fixed distances and high-quality images matched to standardized photos, AI will exceed what any human can do," Dr. Dunn says.

"Right now, when the conditions are less ideal, humans can still have an advantage - especially with people we know well - because we bring context and familiarity to the task. But that gap is narrowing as AI evolves."

Implications

The researchers say the study offers insights into human visual expertise and could inspire improvements in facial recognition technology.

"It shows face recognition skill isn't just about what happens in the brain later, it starts with how we look. The way we explore a face shapes what we learn about it," says Dr. Dunn.