OCR crashing?! What the cure?

Report · Feb 05, 2017

I'm subscribed to Acrobat Pro DC (continuous) on my Mac and I am having a terrible time when trying to OCR documents. Recently I attempted to OCR a 250+ page of printed typed text. Regardless of the settings used, Acrobat would crash during OCR.

I divided the document into parts and attempted OCR scanning of each part. Some were successfully scanned. By a lengthy process of dividing and scanning, I was able to identify the page that appears to be causing the trouble. The question now is what to do with it. Is anyone at Adobe even interested in trying to analyze Acrobat's deficiencies?

Accurate, reliable, and fast OCR is a very important function in my work. It seems Acrobat's abilities in this area have been declining. I find it's becoming harder to use and less informative with its help files, manual, and reduced abilities for user customization. Eg., would it be impossible to place a "Search" bar within the Preferences menus? It would be very helpful.

Thanks.

Report · Feb 05, 2017

Hi,

We are sorry for the issue you have been facing and appreciate your effort in isolating the troublesome pages for us to analyse.

OCR is indeed a very complex and important feature and we'd be more than happy to help you in the issue.

Even though you have mentioned that settings doesn't matter for the crash but still we'd like to know some steps for reproducing the issue.You can share the faulty pages and crash logs via Adobe Send.

To create logs please follow these steps.

1. Retry the scenario where Acrobat crashes

2. Do not close the pop-up when Acrobat crash

3. Open Task manager

4. Right click on “Acrobat.exe *32” process and select ‘Create Dump file’

5. It will create a log file in temp folder. (Path will be displayed on pop-up comes when logs created).

You can also check this under:

a. Open Run process (Win+R)

b. Type %temp%

c. Logs file will be available here with name Acrobat.DMP

Steps to share the file using Adobe send

Share the file using https://cloud.acrobat.com/send

a. Open this link

b. Click on “Select files to Send”

c. Click link "Select file from my computer" and Select the file

d. Click on Create link

e. Share this link

Thanks

Report · Feb 06, 2017

Thank you for your help. I should have mentioned I'm running Adobe Acrobat Pro DC (continuous):

Architecture: x86_64

Build: 15.23.20056.213124

AGM: 4.30.67

CoolType: 5.14.5

JP2K: 1.2.2.37658

If you can provide me with corresponding instructions for Mac OSX, I will be glad to provide the logs and other information you need. Meanwhile, I will see if I can share the faulty documents via Adobe Send. Thanks.

I have uploaded the original intact document, along with the one page I found that seems to always crash during OCR. Here is a shared link to the files:

Shared Files - Acrobat.com

Report · Feb 07, 2017

Hi Cuba_D,

Got the files, they were already OCR-ed i.e., the text was select-able and hence thye may not behave the same way they behave at your end.Please send the non-ocred version of the files. However i still ran OCR on them With Language English and no crash was observed.

For mac, whenever Acrobat Crashes there comes a dialog which presents the detailed crash logs.

You just need to copy them and then paste and save them in a text file and send it to us.

Thanks

Report · Feb 20, 2017

Hi risharmaAdobe,

Apologies for my delayed reply. As you requested, I am sending along an un-OCR'd, untouched original of the document that crashes. I am also sending along one (1) page from the document, that I have identified as crashing Acrobat every time I attempt to run OCR on it.

I am also uploading a text file containing the Crash Log report.

Thanks for your help on this. OCR is a very valuable tool for me, and one of the primary reasons I went with Acrobat. I only wish that Acrobat would include the capability of OCR'ing rotated pages, as is available in the Windows (but not Mac) version of BlueBeam's Revu Extreme.

Here are links to the uploaded files.

Shared Files - Acrobat.com

I hope I got that all correct. I look forward to your response. Thank.

Best regards,

Henry

Report · Mar 03, 2017

Hi ReckonerX,

This seems to be the same user as Cuba_D.

Please confirm the same.

Also to add, i have gone through the files provided by you and logged a bug for the same.

Please also share the settings and steps you used for OCR.

Edit:

The issues is being investigated and we will update you once it is fixed.Stay tuned.

Thanks,

Rishabh

Report · Mar 26, 2017

Hi risharmaAdobe,

Thank you for looking into the OCR matter. BTW, you are correct re. Cuba_D. It is a user account on my old MacBook that I still use at times.

Regarding the OCR steps and settings, I imagine they are rather standard. Depending on the number of documents at hand, I will use "Enhance Scans/" and "Recognize Text In This File" (or "In Multiple Files"). I select the settings for "All Pages", English language, Output as "Editable Text and Images", and Downsample To "300 dpi" (sometimes 600 dpi or 150 dpi if I encounter issues.

These steps and settings work fairly well most of the time. I have been testing ReadIris and considering Bluebeam Revu for the ability to OCR both horizontal and vertical text, but that process is on hold until I find more free time.

The documents I sent previously proved to be really frustrating. It seemed to me that something on just a few pages was crashing the process. I hope you are able to make something of it. Thanks for your assistance.