Optical character recognition scans image-based files looking for text and tries to recognize individual characters. Once visual clues inside the document are matched with any character in the underlying character database, OCR produces machine-encoded text that users can edit in word processors.