New Scientist has recently revealed in a report that researchers at Duke University demonstrated the `InSight' app - the first, somewhat creepy, app for Google Glass - at a technology conference in Georgia last week.

The `InSight' app - which has partly been funded by Google - can help the users of Google's high-tech glasses to pick out their friends in a crowd by recognizing them from the clothes they have worn, even if their backs are turned.

According to the details shared in the research paper of the InSight app, the system essentially helps create a "self-fingerprint" of the user's clothing with "a few opportunistic pictures" that their friends have clicked, using a companion app on their smartphones, when the user is checking e-mail, surfing the Internet, or doing other such things.

With the help of the Internet or Bluetooth, the "self-fingerprint" of the person's clothing can then be shared with Google Glass users; thereby facilitating the high-tech glasses in matching the attire that is seen against the InSight app's database of self-fingerprints.

For improving the accuracy in the identification of friends via InSight's "self-fingerprints," the research paper also describes some other possible methods, like clicking a short video to determine the "motion vector" of a person. Moreover, once the InSight system recognizes someone, it can add more information to self-fingerprints, and can thus pick out a person in a crowd even more easily.