Pattern Is Prologue: Organic Facial Recognition Software Of The Humankind

They never forget a face. Research delves into how ‘super-recognizers’ can do this.

Super-recognizers never forget a face. They need to focus on it only once to instantly recognize it again, even if they encounter it years later, and sometimes even if they see only one feature, such as the eyes. 

They also can get a pretty good idea of what a face looks like in profile if they initially see it straight on. Most ordinary people see faces differently. Their brains take a frontal facial snapshot, which usually is how they remember it — if they can remember it at all. 

“Super-recognizers definitely have an extraordinary skill that scientists have only been investigating in the last few years,” says Josh Davis, professor in applied psychologyat the University of Greenwich in London, and a super-recognizer expert. “We are only just learning about how they do it.” 

Super-recognizers belong to an elite group — experts estimate their numbers at less than 2 percent of the population — and are at the top end of a facial cognition spectrum that includes prosopagnosia, or “face blindness,” or exceptionally poor facial recognition, at the bottom. 

“Society operates under the assumption that everybody is about the same at recognizing faces, and that everyone sees the world in the same way,” says Richard Russell, professor of psychology at Gettysburg College, and an author of a 2009 paper that first described the existence of super-recognizers. “That simply is not true.” 

Experts are studying super-recognizers, trying to learn more about the science behind their unusual ability. Studies in identical twins and among parents and their children suggest it is heritable, although researchers have not yet found a genetic component. 

Also, the brain activity of super-recognizers differs from ordinary people. They show greater electrophysiological activity when processing recognition, meaning their brains show bursts of electrical activity sometimes stronger than that of controls. 

“Shortly after seeing a face they recognize, the scan will show spikes of brain activity not seen in average people,” Davis says. The skill is so brain-dependent that one super-recognizer, after suffering a debilitating stroke, lost his super-recognizing ability. 

“My colleagues tested him again, and he then tested average,” Davis says. 

Most super-recognizers can recall a face even years later — and even if they’ve seen it only once and briefly — whether in person or in a photo. They can remember children’s faces as well as faces across racial and ethnic lines different from their own. Moreover, during this pandemic time of widespread mask-wearing, they can recognize people from their eyes alone. 

“For people we know well and have seen under various conditions, our brains have rich representations of their faces stored in memory,” says Meike Ramon, a cognitive neuroscientist and an assistant professor at the University of Fribourg in Switzerland, who also heads the Applied Face Cognition Lab, funded by the Swiss National Science Foundation to study super-recognizers. “For unfamiliar faces, with no such robust memory representations available, we have to rely on the information we see. For a never-before seen face shown as an image, the average person derives a snapshot-like representation, while super-recognizers automatically get an idea of what that person’s face looks like from other angles.” 

Experts believe super-recognizers also probably can see a child’s photo and identify the adult the child later becomes. Russell’s paper describes a “before they were famous test,” and other anecdotal evidence further supports the idea. 

“In my lectures and demonstrations I use childhood photos of Britney, Angelina and Brad aged from about 3 to 8 years,” Davis says. “Super-recognizers are so much better and quicker at it. If the audience is mainly super-recognizers, a sea of enthusiastic hands goes up, all bursting to deliver the correct answer. If I show the same photos elsewhere to an audience of the same age and interests, you can sometimes hear a pin drop until I offer them a series of clues.” 

Regardless of the scientific underpinnings, super-recognizers have become useful in police and security work where suspect identifications and face matching — from passports, for example — can be crucial. 

Andrew Pope, 43, is a super-recognizer with the West Midlands Police, Great Britain’s second largest police force, where he works as a police community support officer in Birmingham for the Safer Travel Partnership, a team that patrols trains, buses and trolley cars with the goal of keeping passengers and staff safe. He has identified more than 2,000 suspects since 2012 after testing confirmed him as a super-recognizer. 

He joined the department in 2005, but he had no idea he was different from his colleagues until his supervisor — amazed at the large number of accurate identifications Pope was making — suggested he be screened. (Members of the public can try a test found at to see if they might be one.) 

“It felt a bit strange, to be honest, as the word ‘super’ is quite overwhelming,” Pope says. “Touch wood, it hasn’t proved troublesome yet; however, it makes it very hard to switch off when I’m off duty. I’m constantly looking at faces when out with family and friends.” 

As soon as he saw the mug shot of a suspect being held on drug possession, for example, he recognized him instantly as the same man captured on camera two years earlier exposing himself on a city bus. “We had been looking for him on that bus route, and as soon as I saw him, I said, ‘That’s the guy from the bus two years ago,’ ” Pope says, adding that the man was charged with a sex crime and pleaded guilty. “It always makes it special when I can find those kind of people because what he did was not very nice.” 

Davis initially was skeptical about the performance of super-recognizers when asked by the London metropolitan police in 2010 to study some of its officers. 

“I was dubious about the idea of some police being better at face recognition than anyone else,” he recalls. “Most of my previous research and teaching had been investigating or discussing ways of reducing miscarriages of justice of all types, including police and eyewitness identification.” But when it comes to super-recognizers, he adds, “I was very wrong.” 

Davis and others have developed a battery of tests to better identify them. Ramon, for example, has been working with the Berlin police to design toolsto help identify super-recognizers among the city’s 18,000 police officers. 

“I think that there are individuals with exceptional abilities, and my work is devoted to identifying these individuals,” she says. 

“Considering their documented performance across very difficult tasks of facial identity processing, I do believe that super-recognizers could improve policing processes,” provided they are identified and deployed properly, she says. 

Yet, despite their growing popularity outside the United States, they haven’t yet caught on here. 

“I really do think U.S. police are seriously not utilizing the skills of their workforce properly,” says Davis, who consults with police departments and government agencies on the use of super-recognizers, including in Australia, Germany, the Netherlands and Singapore. He says he has had numerous conversations with senior police officials in large American cities — he won’t name them — at their invitation, “that went nowhere,” he says. “I failed to convince them.” 

Stephanie Gessner, an official with the National Association of Police Organizations, a coalition of U.S. police unions and law enforcement associations, responding to an email inquiry, said the group had no information about super-recognizers in this country. 

Davis speculates that U.S. police departments prefer using technology — computer facial recognition software — than humans. 

“Their mind-set is ‘if a computer can do this job, why do I need a human?’ ” he says. 

Yet, he points out that research suggests a combination of the two is more accurate than either alone. 

“Humans see faces in a human way, computers see faces in a computer way,” he says. “They aren’t working the same way. There are occasions when humans make mistakes on faces that no computer would, and occasions when computers make mistakes on faces that no human would. If a computer makes a decision and a super-recognizer looks, chances are they will make the same decision. If you combine the two, you get the highest levels of accuracy.” 

Super-recognizer skills proved invaluable during riots that erupted in London and other English cities for five days in August 2011 — better than software. 

The violence was prompted by the police shooting death of a 29-year-old during a police intervention targeting gun crime. In the aftermath, about 20 super-recognizers identified an estimated 600 suspects responsible for the destruction by sifting through thousands of closed circuit television (CCTV) images, Davis says. 

One super-recognizer, Gary Collins, an East London constable and local gang specialist, recognized nearly a third of the 600 rioters identified. He even pinpointed Stephen Prince, a known offender who was looting, torching cars and attacking bystanders, and whose face was covered by a bandanna, leaving only his eyes visible. Collins, who hadn’t seen him for several years, still nailed him. Prince confessed and received a prison sentence of over six years. 

Interestingly, police facial recognition software identified only one rioter, “and a super-recognizer had already identified that person,” Davis says. 

In the security field, research also suggests that super-recognizers outperform non-super-recognizers in verifying identities — in airport settings, for example, in matching people with their passport photos — even though this skill relies on facial recognition rather than memory. 

“Super-recognizers are much more accurate,” Davis says. “No human is 100 percent accurate, but on average they make far fewer mistakes than people with average ability.” 

Andrew Pope was unaware of his talent growing up in Redditch, a town south of Birmingham, where he still lives, although he noticed he could watch TV actors and recall having seen them in other programs years earlier. 

“I thought everybody could do it,” he says. 

Usually, he will see someone in person — on a bus or in the street — then match the face on CCTV if they commit a crime, although it also happens in reverse. “I do also view CCTV stills of offenders I don’t recognize and have then been able to identify them on the street having come across them when on patrol,” he says. 

He says his parents don’t seem to be super-recognizers. But he thinks Freddy, his 12-year-old son, might be. The child has begun to show similar skills, and he says he wants to become a detective. That would be fine with his dad. 

“I love what I do,” Pope says. “The fact that I can use my ability to help identify suspects and bring offenders to justice is a feeling that is indescribable.” Federal government to expand use of facial recognition despite growing concerns Why we often remember the bad better than the good  The amazing skills of memory athletes© West Midlands PoliceAndrew Pope, 43, is a police community support officer with the West Midlands Police, Great Britain’s second largest police force. He has identified more than 2,000 suspects since 2012 after testing confirmed him as a super-recognizer.

Microsoft and partners may be compensated if you purchase something through recommended links in this article.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s