Does anyone know a way to take an image captured on the iphone's camera, and do some image processing (e.g. edge detection, skeletization), and then overlay parts of the processed image on the original image (e.g. only the highlighted edges).
More generically how do I create a UImage with transparency (do I just scale the image and overlay it with an alpha value, does UIImage support transparency like gifs do). I'm thinking that you could combine a UIImagePickerController with a background thread that takes "screenshots" of the UIImagePickerController view and does image processing on it to detect various objects and provide an overlay augmented reality display.
There's an open source simple image processing library for the iphone. The demo shows an example of taking an original photo (of a sudoku board) and then overlaying the detected object in the original photo.
They explain some of the high-level techniques on their blog .