One of the first apps developed specifically for Google Glass, InSight can pick a user's friends out of a crowd, even when their faces are hidden.
An app that can locate your friends in a crowd based on what they're wearing is a potential use for a headset like Google Glass.
Co-funded by Google and developed by Srihari Nelakuditi at the University of South Carolina, in partnership with Romit Roy Choudhury and colleagues at Duke University in Durham, North Carolina, InSight identifies a person from their "fashion fingerprint."
Facial recognition systems already exist, and anyone with a Facebook profile is reminded how accurate they can be each time they upload photos to the site, but facial recognition systems only work when the face in question is looking directly at the camera.
InSight records the patterns, textures and colors of clothing that someone is wearing and uses that information to make an identification so that when someone wearing a Google Glass headset looks around them, the app picks out that person's friends, and shows their names in the headset's screen.
As the developers explain in an academic paper to support the technology: "Our technique is complementary to face recognition, and exploits the intuition that colors of clothes, decorations, and even human motion patterns, can together make up a 'fingerprint.' When leveraged systematically, it may be feasible to recognize individuals with reasonable consistency."
Google claims that since developers got hold of the Glass headset over 80 new uses had been found for the technology, but due to confidentiality agreements, no information about these apps or uses can be made public. Therefore InSight represents a first in that it is the first potential use for the headset to reach the public domain.
In tests with 15 volunteers, its developers claim that it is 93 percent accurate. However, there are two potential stumbling blocks. The first is that InSight needs to record information from a corresponding smartphone in order to work. The app uses the handset's forward-facing camera to take a series of photos of users while they are using their phones and uses those images to create what the developers call a 'spatiogram' (the colors, textures and patterns of their outfit). In his TED talk on February 28, Google co-founder Sergey Brin took to the stage, sporting a Google headset and proclaimed that smartphones were "emasculating" and that Google Glass was the future. If everyone is wearing Google Glass, which has no user-facing camera -- no one can provide the data that InSight needs in order to work.
The second stumbling block, of course, is privacy. Are consumers prepared to share their dress sense and fashion sense with a Google product? The developers are quick to stress that the system is temporary. It doesn't store historical wardrobe choices (yet) and therefore is only capable of identifying someone for one day at the most. Therefore the app can only be used for identifying specific people on a specific day or in a specific situation, such as trying to find friends at a concert or an airport or before a big football match.
And therein lies potentially a third problem. Even in a huge crowd, everyone is going to notice the person wearing a Google Glass headset. If you're wearing Google Glass, don't worry, your friends will find you.
Read the academic paper: http://synrg.ee.duke.edu/papers/insight-final.pdf