This app would start by using the smartphone's camera to take an image of someone you encounter at random (you have to point the camera appropriately, of course).
Then it would scan the Internet to identify that person.
Then it would seek out all the available profile information that that person has self-posted, about self. If the app finds any of certain key items, then the the word "loser" is displayed on the screen, along with the image of that person.
The fun part is how you will have previously, when the app is installed in the smartphone, designated that certain profile info items will identify someone as being a "loser".
For example, you might designate all Creationists as losers. So, if the app finds that the person's profile indicates that he or she is a Creationist, then that suffices for the app to display "loser" upon that person's image.
More, the app can look for multiple matching things (smokes? pro-choice? Mafia-connected? etc), each of which by itself--for YOU--would be a "loser" flag, and then apply stress to the displayed word, such as: Loser Loser! LOSER! LOSER! (yellow letters) LOSER! (orange letters) LOSER! (red letters)
There could also be an equivalent "winner" app. Sold separately, of course!-- Vernon, May 20 2012 This seems to be contingent on the assumption that a person can be identified automatically from their photo. Is this, in fact, the case?-- MaxwellBuchanan, May 20 2012 Cheaper to do this with a stack of post-it notes.-- pocmloc, May 20 2012 [MaxwellBuchanan], there is "cloud" software that is working on that problem. For now, this Idea is Half-Baked.
I should mention that one possible way of identifying a photograph involves comparing the image to the image that accompanies a self-posted public profile, as on Facebook or equivalent. If trends continue, in the future just about everyone will have done that...-- Vernon, May 20 2012 // comparing the image to the image that accompanies a self-posted public profile//
I have not kept abreast (which suggests another idea, but we shall let it pass) of facial recognition software, but I believe we are still a very long way from being able to correctly match two photographs of the same face under different circumstances, even when the number of possible faces is quite limited. Given a hundred million facebook images, I'd imagine that your chances of picking the right one would be somewhere between one in a hundred million and one in a million.
In fact, I am pretty sure that if you showed me 100 million faces, I'd find a fair few that looked just like Walter Matthau, or like my granduncle Thaddeus (with or without his monocles).
Still, there's a lot of future out there waiting to happen.-- MaxwellBuchanan, May 20 2012 Not for us sadly.-- rcarty, May 21 2012 The app Recognizr by TAT did the part that isn't identifying losers several years ago.-- notexactly, Mar 31 2016 A local newspaper had (when it still existed) a similar idea: each friday after a food, livestock and groceries market which is scheduled every 2nd thursday they published a picture of a market visitor and declared him or her a winner. To claim their price winners had to come to the town again (loser's count: 1), locate the editor's office (2), identify themselves a the depicted persons (3) and claim their price of (if I remember right) about $20 (4). The newspaper never published if the price was ever claimed.-- Toto Anders, Mar 31 2016 random, halfbakery