how could I get “hold of” two points coming from the iPhone’s camera (in live stream) like these people do it : http://www.robots.ox.ac.uk/~gk/youtube.html (they’re using this technique to bypass the need for markers in AR..)
I’m not interested in AR, I’m only interested in coming up with a way to “hang on” to such points coming from the camera’s live stream and not lose them regardless of whether I’m moving the camera closer to them or further away, to the left or right,..etc.
Is it just a matter of coming up with a code that scans the camera’s input for something that is “standing out” (because of diferrence in color?high contrast, etc?)
Thank you for any helpful ideas or starting points!
Check out http://opencv.willowgarage.com/wiki/
OpenCV is an open source lib that can do lots of things around image recognition and tracking.
If you google for it, along with iOS as a keyword you should run into a few related projects that might help you further.