Acrobat 2017 optimize OCR on watermarked documents

Hi ATP2018,

Let me add to Akanchha statement: Having not seen the previous or current documents with watermarks, it's impossible to know exactly but consider: how could or should any application at this point in time know what's underneath the watermark?

I say "at this point in time" because it's always possible that at some point, machine knowledge (or AI) will recognize full sentences and be able to conjecture what text might be missing. We kinda/sorta see this when auto-correct on our phones assumes it knows what text we meant to write (and you know how accurate that has been up to date).

There is a long-standing observation with Photoshop when someone is teaching and shows how to remove unwanted objects, someone from the audience will ask how does Photoshop know what's behind the sign that was just removed? [Well, PS doesn't know any more than the PS user who cloned something near the sign to replace the sign.]

My point is that any OCR package can only work with what it has. As Akanchha, pointed out, a poor quality scan has a direct affect on the quality of the resultant OCR. For example, the text combination "ir" may be seen as "n" or "ii"unless the scan is of very good quality and high resolution. If there is ANYTHING that obscures or alters the text, be it a watermark or a pen (or pencil line), all bets are off.

Hopefully AI will get better on this but by then, and if Skynet gets any more powerful, all bets are off.