This question comes from thinking about this toy usb camera that my bro got for christmas. One day I wanted to get an image onto my computer from a paper I had--but since I have no scanner, I decided to see if the camera could capture it at all. I showed the paper to the camera and had the computer take a snapshot. Naturally, it was dismal and unreadable. But then a question popped into my head: would it be possible to get a piece of software that could "lock on" to the paper, and then use several frames of video of the paper (being moved slightly) to get a nice, sharp image? I ended up re-typing what I had hoped to scan--but the question stayed with me. I would love to see such a thing in action, if it is possible. Do any of you know whether such a thing already exists?

I don''t know. A movie is merely a bunch of still frames. If the picture you took was crappy, than there''s no indication that any of the other pictures (or frames) will be any better quality. In the end, I''d think you''d just end up duplicating the data instead of refining it. I''m thinking you''re already doing the best you can with the single still frame.

I''m imagining the situation where the overall image is offset on the cameras sensor by some fraction of a pixel, thereby producing more information.

But maybe you are right.

Why don''t you make several photos up close and then stick them together into on final image in an OCR program?