Is there any way to recognize using IEVision the coordinates of a "white box" in the image that could be the borders of a page or the borders or a typical ID Card?
This would be to pre-process with not so high CPU consume the image (aimed to video processing) and only after having found the "white box", applying the OCR or BarCode processing on that sector only.
I saw the Image Alignment sample, but it is only partially the way to get what I'm thinking.
Are you saying it is a document that is not standard (may change from one document to the next)? Are you able to post some examples so we can understand the problem better?
Hi Nigel. Sorry I haven't answered to you previously.
The main idea is to detect if you have somehing as and ID Card (parametrized as a white rectangle with a picture in a specific position, for example), so if you recognize that pattern in a live video stream then you make a more in depth processing with an OCR (which can take a much longer CPU processging) to get all the information on it.