21

I have been searching Google for some time but cannot find an answer to my question.

I have unwanted layers of OCR in a document that I recently scanned with Adobe Acrobat. It has not been OCRed properly, and I want to redact some information, but the OCR is making the wanted information to get erased. I converted the files to TIFs, but noticed a (very) significant quality loss. I have heard that printing to another PDF either keeps the text or reduces the image quality.

1
  • Is there any programmatical way to do it, via python?
    – Mohith7548
    Jun 24, 2021 at 7:29

11 Answers 11

11

In Acrobat Pro DC, the appropriate command is "Remove Hidden Information," which is available through both the "Protect" and "Redact" tools.

On running the command, it just searches out the hidden information but does not change the document. You must then tell Acrobat which information to remove. In this case, select "Hidden Text" in the Results pane, then click the Remove button and save the changed document.

3
  • 2
    I have used the "remove hidden information", but for me for some reason that just removes parts of the image on certain pages. Thanks for your reply however.
    – Sanoo
    Apr 11, 2017 at 4:20
  • This is not universally true. Somehow (probably macOS PDFKit bugs) my ABBYY FineReader-OCRed text got corrupted, and checking "Hidden text" under Redact → Remove Hidden did remove the text without any issues; I was then able to successfully use Enhance Scans → Recognize Text to perform OCR within Acrobat itself. Jan 21, 2018 at 20:16
  • 1
    The problem for me is that after I remove the hidden text, I'm still not able to run an OCR with "ClearScan" (i.e. "Editable Text and Images"). It's strange because the text layer appears to be gone, yet running OCR produces the error "Acrobat could not perform recognition because: page contains renderable text." Sep 18, 2018 at 10:38
4

Try the "MS Print to PDF" driver. It ships with all recent Windows versions. Make sure to check "Print As Image" under advanced settings to remove OCR.

The quality loss in printing to PDF is negligible. It does however keep the OCR by default unless you print as image.

enter image description here

3

After a lot of experimenting, I found that printing to Adobe PDF from Adobe Acrobat prints the document without the OCR and without losing the quality (an unnoticeable at first glance resolution is lost).

However, many sites claim that this does not work. I also tried the other printers such as Foxit Reader and OneNote but the quality was reduced. JPEG too was the same.

Please keep in mind that your mileage may vary.

Note: I am leaving this thread marked as unanswered in hope of finding a better answer than mine.

3

In Acrobat Pro: use 'remove hidden information' (under 'protection'). Select all, execute, OCR is gone

3

If, as you say, the documents are scanned and not printed to PDF from Word for example, you can easily remove with your Adobe:

Select Document, Examine Document and now you can remove the hidden text (OCR).

2
  • Thanks for your reply. I'll test it out as soon as I can and let you know. Thanks for the answer!
    – Sanoo
    Feb 19, 2016 at 14:31
  • 1
    I thought I already commented on this, but the problem is that I have Acrobat DC Pro, and those menus have been removed. Thanks for your answer anyway.
    – Sanoo
    Jul 17, 2016 at 7:43
2

Easy way to remove OCR layer from PDF: open PDF in Firefox and "print" into another PDF.

Note that "nice" PDF (e.g. created by MS Word) will become much larger (in my case, from 0.5 to 2 MB), and quality is reduced somewhat. Make sure you set correct paper size when "printing".

If you want to redo OCR instead of removing it completely, and you don't mind command line, use ocrmypdf:

ocrmypdf --redo-ocr --output-type=pdf input.pdf output.pdf

On Windows 10, the easiest way to setup and use ocrmypdf is via WSL.

1
  • Thanks SO much for this. On MacOS Preview I had used the Redact feature, but the text remains searchable so that "Find" in Prevew helpfully shows surrounding text even if redacted in the visible version. Doing "Print" from Firefox seems to have converted every page in the PDF to an image and Find no longer does anything useful.
    – user31096
    Apr 5, 2022 at 4:42
1

In Acrobat X, under Protection, there is a Sanitize Document button that removes EVERYTHING but what can be seen (including OCR'd text layer), converting the document to a flattened bit map.

1

I solved it by exporting to JPEG, then from JPEG 'combine files in acrobat'. This is from a doc that was originally a word doc and had been converted to PDF. OCR is gone.

0
1

Use the PitStop Pro Acrobat Plug In, in the "Actions List ", create a new action, in the upper right, look for" Select text fragment "and" Remove selected object ", run scope: whole document as seen below:

screenshot

0

I built a tool to do this free PDF Redactor. If you upload the image and just click redact it'll flatten your pdf and remove OCR. If you want you can also draw redaction marks on the document as well.

1
  • "File size should not be more than 5MB"... Really? Most scanned-and-OCR'd PDF files are larger than this.
    – Pertinax
    Oct 19, 2022 at 13:33
0
  • For Adobe X and above: Tools > Protection > Remove Hidden Information.
  • For Adobe 9 and below: Document > Examine Document.

Reference: https://answers.acrobatusers.com/undo-recognize-text-q28083.aspx

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .