How do I improve OCR tool?

[CVHManchester]

I have scanned many pages of a C19th journal.

I then used Acrobat to OCR a section. But the quality is appalling and unreadable.

I have tried 3 different file formats (PDF, JPG, PNG) to drag and drop below an image of the page (all less than 47MB) but they all produced the error message "Sorry, unable to complete the action you requested. " 

[Amal.]

Hi there,

Hope you are doing well and thanks for reaching out and sorry to hear about the trouble.

The OCR quality in Acrobat depends heavily on the scan clarity, resolution, and text contrast. For older documents like 19th-century journals, OCR can struggle if the text is faded, ornate, or scanned at a low resolution. Here are a few steps you can try:

• Check scan resolution – For best OCR results, scans should generally be at least 300 dpi (black and white) or 600 dpi (for grayscale/complex fonts).
  
  
• Clean up the scan – If possible, enhance the image first (adjust brightness/contrast, deskew, or remove background noise). Acrobat’s “Enhance Scans” tool can help https://adobe.ly/3IqsVfV
   
• OCR settings – When running OCR in Acrobat, make sure the language is set correctly to match the text. For historical documents, try “Searchable Image (Exact)” instead of “Editable Text” to preserve readability.
  
  
• Error with file upload – The “Sorry, unable to complete the action you requested” message may occur if the file exceeds system handling limits or if there’s corruption. Try saving the scan as a new PDF through Print to PDF or Optimize PDF in Acrobat before re-running OCR.
  
  

If it still doesn't work, could you send us the PDF file in question so we can check it at our end.

~Amal

[gary_sc]

Hi, @CVHManchester,

First off, as someone who's done OCR for about 30 years, I always marvel at it being possible in the first place. That notwithstanding, the better the original scan is, the better luck you will have when doing OCR. A bad image will NOT help. A good image is better, and the only way to get great results is to start with as good a scan as possible. Just like a photograph, if it's out of focus and you do not have an obvious subject in your photo, Photoshop cannot help.

Some years back I wrote a blog for Adobe on how to improve your scans, here's a link. I hope you get some help out of it.

https://community.adobe.com/t5/adobe-community-professionals/scanning-clean-searchable-pdfs/m-p/4785...