OCR fails in Adobe Scan Premium on iOS (iPad Air 5th gen)

Report · Jul 12, 2024

I'm using premium subscription of Adobe Scan to scan books (less than 100 pages each) on iPad. Adobe Scan is the latest version (24.07.16). iPad Air (5th generation) is running the latest iPadOS 17.5.1. The pad has more than 200 GB of free memory and is plugged to power. I have enabled OCR on Adobe Scan's preferences and selected the right language. Book pages are properly aligned and lit (I'm trained AV professional).

I'm facing two problems:

1) Taking pictures of the pages (using the book mode) works otherwise ok, but the screen blurs and freezes from time to time (approximately every 30 pages or every 5 minutes). When I put the app to background and reopen it, it works again. This is mildly frustrating. Next is worse:

2) When I've taken the pictures and I select Save PDF, Adobe Scan starts to upload the PDF to Adobe Cloud. After uploading succesfully Adobe Scan will display message "OCR..." or "waiting for OCR...". Message stays on for couple of minutes. Then it disappears. OCR fails for some reason, no error messages given. When I open the PDF (either from Adobe Cloud using Acrobat, or from iPad itself) there's no trace of OCR: text is not selectable. When OCR process fails like this, is there a way to force Adobe Scan to retry doing OCR on PDF file?

In maybe 30% of my scans the OCR process works succesfully. In 70% of scans it fails and I'm unable to start it again. This error seems random, as my scanning method is always the same. Pages I scan are text only, no pictures, it really should be easy job for OCR. Please help!

Report · Jul 15, 2024

Hi @Elokuvakoulu,

Hope you are doing well. Thanks for writing in!

It is always advisable to create smaller chunks of files when scanning a bigger file, to later combine them for the complete PDF.

The reason for that being, even though you might have huge amounts of storage available on your device, there is limited memory when it comes to processing power of the device.

Also, mobile devices or iPads have certain background processes running, which piles on to the already draining memory to create the PDF and run OCR on it.

You will also notice a better experience in performing OCR when you create smaller fragments of the file.

For the already created files, you might want to check the Edit Text Feature and see if the app runs OCR successfully on the file.

If not, please share a screen recording of the issue for a better understanding and assistance.

-Souvik

Got your issue resolved? Please label the response as 'Correct Answer' to help your fellow community members find a solution to similar problems.

Report · Jul 18, 2024

Hi,

and thanks for the reply. I find it a bit hard to believe that 90 pages file (file size about 90 MB) would be too much for a fairly new device to handle. How small PDFs would you suggest for me to scan?

Edit Text feature seems to force Adobe Scan to do OCR, thanks for the tip. However it seems to do OCR for only one page at the time (the page which is visible). I'd prefer not to flip through hundreds of pages manually (I've got many less-than-100-page documents waiting for OCR). Is there a way to run or rerun OCR for the whole PDF?

Report · Jul 22, 2024

[Edited Response]

Hi @Elokuvakoulu

Hope you are doing well, and thanks for reaching out.

I checked the Adobe account associated with your Adobe ID and found that you have both a personal and a business account with the same ID.

Your personal account includes Adobe Scan premium services, which allows you an OCR service limit of 100 pages, compared to the 25 pages with the free service. Please ensure you are logged into the personal account to utilize the increased OCR page limit (https://www.adobe.com/devnet-docs/adobescan/ios/en/managingsubscriptions.html#ocr )

For more information, please carefully go through the help page (https://helpx.adobe.com/enterprise/kb/enterprise-id-faq.html)

Also, as mentioned the edit text feature seems to do OCR for only one page at the time and you want to run the OCR for the whole PDF, please make sure the option to 'Run text recognition on saved PDF' is enabled from the App Preferences (Tap on profile icon > Preferences).

By default, the app automatically converts image text in scans to actual text using the Optical Character Recognition (OCR) feature. The default text recognition language is set to the language of your device. Automatic text recognition allows you to save, edit, markup, and do more with text.

For more information please check the help page https://www.adobe.com/devnet-docs/adobescan/ios/en/settings.html#change-text-recognition-settings-oc...

Hope this information will help

~Amal

Report · Jul 31, 2024

Hi. Yes, I'm using an account with premium subscription (as I wrote in original post).

Report · Aug 02, 2024

As mentioned earlier, I have enabled OCR ( 'Run text recognition on saved PDF' ) on Adobe Scan's preferences and selected the right language.

Summary:

When I save PDF, Adobe Scan begins to run OCR, but it either stops before it has finished it OR it gets stuck in the middle of OCR without ever finishing the process. Either way, no OCR data is saved within PDF file.

By using "Edit Text" feature I can rerun OCR for one page at the time, but I want to re-run OCR for the whole PDF. Is there a way to do so? If not, can you add the feature?

P.S. The Edit Text feature's OCR process gets also stuck on certain pages (typically on pages that are empty or almost empty), but at least Edit Text feature lets me press cancel and move to the next page to continue OCR from there.

Based on everything that has happened, could it be that your OCR algorithm does not have a time out -function which would make it move automatically to the next page if it gets stuck on a page it can't figure out?

Report · Aug 02, 2024

Hi @Elokuvakoulu,

Hope you are doing well. Thanks for writing back!

By default, Adobe Scan performs OCR on all pages if the file is within the 100-page limit. If you are not seeing the results of automatic OCR, I suggest checking the app's preferences.

You can go to "Home-> Profile-> Preferences-> toggle the button "Run Text Recognition."

Once done, try scanning another file and let us know if it works.

The other method could be, once the scan is done, you can click the three dots-> Modify Scan-> click the enhance scan icon on the top right to perform OCR.

By design, Adobe Scan performs OCR once the file is uploaded to the cloud. You might observe that the file shows a message "Uploading to Cloud" once it is created. After uploading to the cloud, you will see a "Running OCR" message, which is when the text recognition process takes place.

If these do not match the case, please share a screen recording video describing the entire event so we can investigate this further.

-Souvik