I'm looking for a recommendation of a good OCR tool.
The one that I use (which will remain unnamed) works well for most things, but can't handle the attached file.
I'm trying to extract the number form this image (the score). Can anyone recommend a tool that can do it?
In the past some of our users have had success with Textract and Kleptomania (OCR) engine. http://www.structurise.com/
Quote from: Deana on June 26, 2014, 08:49:00 AM
In the past some of our users have had success with Textract and Kleptomania (OCR) engine. http://www.structurise.com/
I installed Kleptomania just now (as a test), but it seems to be one of those "only works with the systray" programs.
This makes it unusable for me.
Several vendors provide dotNet OCR APIs. Since WinBatch now hosts the CLR there is a good chance that one or more of them work with WinBatch. I have never tried any of them so I cannot make a recommendation nor guarantee success. And, of course, it would require a recent version of WinBatch...
There is always MODI, but that requires Office 2007.
FYI/FWIW, Textract (the command line version of Kleptomania) couldn't decode my file either. Just rendered it as a ":" character.
As you can well imagine, none of the other options suggested in so far in this thread are available to me, so I've about given up on this. Still, I'd be curious if any of them would work on my file; if anyone feels inclined to try - to take my file and see if your pet OCR engine can decode it, I'd appreciate it greatly.
Quote from: snowsnowsnow on June 28, 2014, 08:31:15 AM
FYI/FWIW, Textract (the command line version of Kleptomania) couldn't decode my file either. Just rendered it as a ":" character.
As you can well imagine, none of the other options suggested in so far in this thread are available to me, so I've about given up on this. Still, I'd be curious if any of them would work on my file; if anyone feels inclined to try - to take my file and see if your pet OCR engine can decode it, I'd appreciate it greatly.
Have you tried Capture2Text
http://capture2text.sourceforge.net/
If I scoped the whole of your screenshot it managed all of the number except the last digit.
JW
Quote from: snowsnowsnow on June 28, 2014, 08:31:15 AM
if anyone feels inclined to try - to take my file and see if your pet OCR engine can decode it, I'd appreciate it greatly.
I took your file, pasted it into OneNote, right-clicked and selected copy text from image and got
Score
-11656
OCR/MODI is still available as a free MS download but is no longer included with Office after 2007. I haven't messed with OneNote that much, but this was cool. Glad I tried it.
That's cool.
Can you remind me what this "OneNote" is? It is the same as MODI (whatever that is) ?
Quote from: snowsnowsnow on July 01, 2014, 05:28:01 PM
That's cool.
Can you remind me what this "OneNote" is? It is the same as MODI (whatever that is) ?
OneNote comes as part of Office 2010/2013. MODI is the Office OCR object (I think it is an .OCX). MODI was included with 2007 but dropped thereafter. I had 2 people at work that I showed how OneNote worked with your jpeg. One's wife is a teacher and has been looking for an OCR solution for on-line lesson plans that have a lot of bitmaps. They were really impressed. (sorry for the segway)
So, I assume OneNote inherited and extended the OCR capacities of MODI. Maybe Deana can test your jpeg with OneNote and confirm that it works as well as it does.
Honestly I never work with OneNote.... But I inserted the image in OneNote 2010, right-clicked on the image and selected Copy Text From Picture. It was unsuccessful reading the test from the image.
Results:
I So ââ,¬Ëœ
ââ,¬Ëœ1
Quote from: Deana on July 02, 2014, 08:53:23 AM
Honestly I never work with OneNote.... But I inserted the image in OneNote 2010, right-clicked on the image and selected Copy Text From Picture. It was unsuccessful reading the test from the image.
Yeah, I just confirmed that. My original test was with Office 2013, but I just tested at work with 2010 and got the same [dismal] results. Way to go Microsoft!!! Apologies to Snow++ if my earlier post was misleading.