clock menu more-arrow no yes

Filed under:

InSight app uses Google Glass to identify people by their clothes

New, 80 comments

If you buy something from a Verge link, Vox Media may earn a commission. See our ethics statement.

Google Glass (STOCK)
Google Glass (STOCK)

Google's Project Glass is fertile grounds for imagining a future where everything can be captured and analyzed as soon as it's seen, and privacy concerns have centered around things like facial recognition. But what if Glass could recognize your friends by their clothes? InSight, developed by USC researcher Srihari Nelakuditi with support from Google, can create temporary "visual fingerprints" based on the color and pattern of a person's clothes. After snapping a series of pictures and matching them to a name, its "spatiogram" can work in circumstances that wouldn't be possible with facial recognition: finding people who are far away or have turned their backs, for instance.

Obviously, the system only works until someone changes clothes or puts on a coat. In a test of 15 people, it was able to recognize them 93 percent of the time. It's also not integrated directly into Glass: it's a smartphone app that connects to the camera and display functions, like most other Glass services at this point. New Scientist reports that its creators envision it being used to find friends in a crowded airport or shopping center; we don't know how long it takes to identify a person, but it could also work for remembering the name of someone you've just met at an event or party. As more developers work on integrating apps into Glass, we expect to see other, similar tools before launch.