(and the OCR part could probably easily be matched with the source code to get perfect error correction)
(and the OCR part could probably easily be matched with the source code to get perfect error correction)