Skip to main content

Why You're Able to Spot a Friend in a Crowd, Even When You Can't See Their Face

And how new research on this common phenomenon will be used to improve controversial facial-recognition technology.


Have you ever surprised yourself by correctly recognizing a friend in a crowd, far, far away? Even if her face isn’t at all visible, there’s something about the way she’s standing or walking that gives her away instantly. New research by psychologists at the University of Texas-Dallas helps to confirm and explain that very common phenomenon.

In a study published recently in the journal Psychological Science, researchers asked participants to look at photographs of people in different settings and clothing and to match them up, determining which ones were photos of the same person. The photographs in the experiment were specifically chosen because they were hard for a facial-recognition computer program to identify. The study found that the participants had more success when they could see the bodies of the people in the photographs—and less success when the bodies were less visible but the faces were clearer. In fact, the participants also accurately matched the photos when they could only see the people’s bodies, while the faces were completely blocked out.

"Our work shows that the body can be surprisingly useful for identification, especially when the face fails to provide the necessary identity information."

What’s most interesting about the experiment is that the participants tended to believe, incorrectly, that they were primarily looking at the faces in the photos in order to identify the people they were looking at. Eye-tracking equipment revealed that that wasn’t the case: when they couldn’t get enough information from the faces in the photos, they were actually spending more time studying the people’s bodies. “People’s recognition strategies were inaccessible to their conscious awareness,” said lead researcher Allyson Rice.

Many psychologists and computer scientists have historically studied facial recognition and ignored the larger picture, the authors concluded. They added that there are clear implications here for the technological development of non-human recognizers of faces, as well.

“Given the widespread use of face recognition systems in security settings, it is important for these systems to make use of all potentially helpful information,” said co-author Alice O’Toole. “Our work shows that the body can be surprisingly useful for identification, especially when the face fails to provide the necessary identity information.”

According to the University of Texas, this study was funded by the Technical Support Working Group of the Department of Defense, and the work of one of its co-authors, P. Jonathon Philips, is partially funded by the FBI. The fact that these agencies are working to find ways to make facial-recognition technology even more accurate than it already is won’t be surprising news, but neither will it be welcome news, to many people.

Facial-recognition technology in general has proved to be a fairly controversial issue, from the ethically problematic crime-fighting and crime-solving tactics pairing the tool with ubiquitous surveillance cameras, to the slightly more banal (but still creepy) example of Facebook photo-tagging. As it turns out, people don’t like to feel as though they’re walking around constantly being scanned through a series of Terminator’s-eye lenses.

Many legislators and privacy advocacy groups have spoken out against the spread this technology, and the debate will surely continue as more stories come out about the sometimes surprising ways that government agencies and corporations have been adopting it. And, of course, there’s the commercial exploitation of that creeped-out-feeling. For instance, there’s a new line of “Glamoflage” T-shirts that are meant to be worn as a kind of facial-recognition-tech-camouflage. (But, unfortunately, they are absolutely hideous.)

Maybe facial-recognition technology could avoid controversy if it weren’t used in massive government databases or for the purpose of coercing us into becoming even more absorbed in online social media. Maybe it could stick to its most appropriate use, in a machine that recognizes when we’re yawning and gives us free coffee. That’s something that everyone can get behind.