OCR text can sometime cannot read character from source, so it will be replaced with question mark "?"
for example, apple can be translated as a??le by OCR because two p characters are unreadable.
for example, aaaaaaaaaaaaa can be ?????????????
but to save space it is interpret as a13

given two ocr texts, compare if they COULD be equal
a2le and app?e are equal

I have done naive way, and make both text a**le, and app*e, and done loop comparison
if string is a23423423bc23 it will not be efficient.