sorry for not making it clear - the bitmaps all ready had text inside of them so we were basically looking for OCR.