by Varun Raj, 6 months ago
The US Patent and Trademark Office has just published a new patent application by Google related to its Project Glass. The patent hints at the many possible features of Google’s futuristic head mounted device. Among…
A new Google-funded app will let Glass recognize and point people in a crowd, based on the clothes and accessories they wear.
When you think about identifying a person, the first technology that comes to mind is facial recognition. But recognizing someone by their faces isn’t always feasible. In low light, in crowded areas, or simply when someone stands with their back against you, that often proves hard or even impossible.
Try finding some friends in a crowded shopping mall. More likely, you’ll recognize them by their clothes and gait, rather than their faces. That’s precisely the idea behind a new app, called InSight, that will let the Google Glass identify and point to people, based on their outfits.
InSight is being developed by students in North Carolina, with financial help from Google. The app takes a unique “fashion fingerprint” of a person, by analyzing their clothes and accessories. The system builds a “spatiogram” that contains the “spatial distribution of colors, textures and patterns” in a person’s outfit. In other words, InSight learns what you are wearing.
Running the app, Glass can compare the spatiograms of people in its field of view with the spatiogram of the person you are looking for. When a match is made, the wearable computer will display the name of the person over the area where the person is located.
Early tests have demonstrated a 93 percent success rate for the system.
From my understanding, the system depends on the consent of the user who will be identified. InSight will snap pics of the user while he or she uses the phone, and build their spatiogram file in the background. Than the person can choose to send the file to another person, who can then load it up on the Google Glass.
Because the spatiogram changes each time you change your shirt or put a new scarf on, it could be relatively easy to evade tracking.
I do think that similar systems will be eventually be deployed by police, security teams, and similar personnel. Just imagine how easy it will eventually be to track people in a crowded train station, based on a quick scan made through a security camera. Till then, there are more benign applications for the app. For instance people suffering from a neurological condition that prevents them from seeing faces are likely to embrace this tool.
The Glass is slowly turning our sci-fi fantasies into reality. Terrifying? Probably. Exciting? Absolutely!