Troubleshooting Document Processing scans, including OCR

The scan action the user selected is meant to create a text-searchable document with OCR, so why is the scanned file not searchable or not accurately detecting the text?

OCR technology is impressive, but not perfect. A failure is usually for one of the following reasons:

• The document has been loaded into the MFD in the wrong position. PaperCut MF attempts to detect the page orientation, however this is not always accurate.

• The DPI of the scan job is too low. Try scanning again at a higher DPI.

• The fonts are complex or artistic.

• The image quality of the original file is poor, for example, the page is damaged or skewed, the page has a lot of speckling, or there isn’t much contrast between the text and the background.

• The language used in the document has not been enabled for OCR on the Capture tab of the Admin web interface.

• The Scan Action does not have OCR enabled for the selected file type (PDF).

Turning on the Despeckle and Deskew options can help improve OCR detection accuracy by enhancing image quality. This can result in slower performance, depending on your infrastructure.

If your document is not private in nature, or you are able to reproduce the problem with another non-private document, send the document to PaperCut support to help us improve OCR accuracy.

I have set up a locally hosted Document Processing server and I get an error message on the Capture page.

• Ensure that the OCR service is functioning and can be reached from the PaperCut MF Application Server.

• If you are using a firewall other than the Windows Firewall (which is configured automatically by the installer), open port 9181 (inbound) to allow connections from the PaperCut MF Application Server.

• On the Capture page, make sure there are no typos in the Document Processing server hostname or IP address.

• Wait for a minute and then refresh the page to see if the error message goes away.

• Uninstall the Document Processing service and install it again.

• Investigate any potential network problems.

When using Batch Splitting, the output documents aren’t separated correctly.

• Ensure the correct duplex mode is set in the scan action Input settings.

• Ensure the number of pages for each document in the input batch matches the number of pages for each document set in the scan action.

• Check that the automatic document feeder isn’t grabbing multiple pages at a time.

With Blank Page Removal turned on, my documents are still being delivered containing blank pages.

• Blank Page Removal works at the scan action level, not globally. Make sure it’s enabled for the specific scan action used to produce the documents..

• Any content on the page will cause it to be classified as not-blank. This includes:.

  • headers and footers
  • page numbers
  • text like “this page left intentionally blank”
  • any image content
  • logo
  • watermarks
  • smudges or shadows.

• You can manually The blank page detection threshold can be manually tuned to reduce sensitivity using [THIS CONFIG KEY] ..

A user’s document hasn’t arrived. What should I do?

If the user received a scan success email:

• Ask the user to check that the destination they are looking in is correct.

If the user did not receive a scan success email:

• Check the PaperCut MF Admin interface Logs > Job Log to see if the job is still in progress.

• If it is, give it a few more minutes to come through. The larger and more complex the document, the longer it takes to process, particularly if Document Processing enhancements like OCR are enabled. For example, a one-page black and white scan at 300 DPI should be delivered to the destination in less than a minute. However, a 20 page color scan with OCR, Blank Page Removal and Deskew at 600 DPI will take approximately 20 mins, depending on the speed of your network.

• If the job has failed, check the Application Log for messages about the problem:

  1. Select Logs > Application Log.

     The Application Log page is displayed.

If you can’t resolve the problem, then contact your PaperCut MF reseller with your Support ID.

The full error message is as follows:
The OCR process has timed out after {0} minutes while converting the scan job by {1} on {2} to a searchable document. Refer to the Integrated Scanning Troubleshooting section in the PaperCut manual.

Follow these steps to adjust the timeout setting.

1. In the PaperCut MF Admin interface, navigate to Options > General.
2. In the Actions menu, click Config editor (advanced).
3. Search for system.scan.ocr.processing-timeout-mins
4. Change the value to the number of minutes (default of 30) after which the job will timeout. We recommend 60.
5. Click Update.
6. Restart the PaperCut Application Server service.

Your OCR jobs will now timeout after the number of minutes you just set.