I read reports here and here last week of two facial recognition startups - Mugr and Eyealike. The consensus in the blogosphere seemed to be that the demo applications put out by both companies were frivolous. Here is an idea for an application that may be just as frivolous but that could potentially do very well on a platform like Facebook, or perhaps even across several networking sites using Open Social.

The Facebook app could be named My Look Alikes or similar and would behave as follows:

1. User installs app.
2. App runs through FB public profile photos of people that are of the same gender as the user and returns a list of people that look like the user. Given the large number of FB members, this may prove too demanding for the app. Therefore, it may be a good idea to run through a subset at a time - limited by number (e.g. 50,000) or network, or geographic location.
3. User rates each suggested Look Alike based on whether he or she agrees with the suggestion of resemblance.
4. User invites friends to rate their Look Alikes.
5. User has option to subscribe to a RSS-style feed of Look Alikes as new profiles get created on Facebook.
6. User gets a “View/Suggest Look Alike” button under his or her profile photo. When clicked by other FB members, such members will be able to view the user’s Look Alikes or suggest one or more of their friends as a look alike of the user. So if I am a user of the app, anyone (user of the app or not) seeing my profile photo would be able tell me about a friend of theirs that they think I look like. These human-suggested look alikes would be listed alongside the app-suggested ones on my “My Look Alikes” canvas page and I would be able to rate them also, and invite friends to do the same.

I am not entirely sure how the app would make money though, except perhaps for the now popular approach of promoting other applications for a fee. So there you go Mugr and Eyealike - a facebook app that you could develop if you were looking for one.