return the position of each recognized word

Nov 13, 2009 at 8:54 PM

Hi Max!

Is is possible to get the position of every recognized word in the original image i.e. not just the output text file?

Regards, Dmitry

Coordinator
Nov 15, 2009 at 6:31 PM

Hello Dmitry,

What you want is in fact maintaning DOM  for image being recognized (text blocks, lines, words, characters etc.) and relating DOM objects to text literlas recognized. CuneiForm recognition engine has such a capability but Puma.NET doesn't use it. It's possible that latter I'll add DOM support into Puma.NET but now there's no such a capability.

Thanks for your interest to the project,

Maxim.

Nov 15, 2009 at 8:23 PM
Hi Max,
I can sponsor the work to the certain degree. all code updates can become public, i.e. i do not need the exclusive rights.. what you think? is the small compensation of 200$ would be enough?
the generated DOM should be better be in XML form.
Best, Dmitry
----- Original Message -----
From: [email removed]
To: [email removed]
Sent: Sunday, November 15, 2009 9:31 PM
Subject: Re: return the position of each recognized word [pumanet:75074]

From: MaximSaplin

Hello Dmitry,

What you want is in fact maintaning DOM for image being recognized (text blocks, lines, words, characters etc.) and relating DOM objects to text literlas recognized. CuneiForm recognition engine has such a capability but Puma.NET doesn't use it. It's possible that latter I'll add DOM support into Puma.NET but now there's no such a capability.

Thanks for your interest to the project,

Maxim.

Jan 18, 2010 at 9:55 AM

Hi Max,

how is the status of this enhancement - i also would be interested on this feature - this feature would be important to generate "searchable" PDF´s which contains the text information in background of the image. Without the position information of the Text this is not be possible. We also would donate some money to get this feature implemented.

best regards

Wolfgang

Coordinator
Jan 20, 2010 at 8:06 AM

Hi Wolfgang,

To be honest I'd spent some time on researching this feature and still didn't find any fast solution. I'm loaded now and don't plan any large activities on Puma.NET soon. So I don't want to guarantee you any terms but if I implement this feature I'll publish new release and write to these branch.

Thank you.

Jan 20, 2010 at 10:01 AM

Ok, thank you - hope you will find a solution and time to implement this important feature.

 

 

Mar 31, 2010 at 4:37 AM

Greetings Maxim,

I'd also like to find text positions and would be very eager to see this feature, if and when it can implemented.

Kind Regards,

Jonathan

Coordinator
Mar 31, 2010 at 1:57 PM

Dear Jonathan,

To tell the truth the answer is the same - don't know when and whether. I have absolutely no time for any activity on Puma.NET and until someone else takes part in development no changes will appear in the near future. 

Best regards,

Maxim.

 

May 21, 2010 at 7:59 AM
Edited May 21, 2010 at 8:00 AM

Hi all,

I have written a piece of C# code which allows to read a Bitmap with Puma directly in memory and retrieve each paragraph, line, word, character along with its position, confidence, characteristics, etc...

If you are still interested in this, you can contact me.

Best regards,

Erik

Coordinator
May 21, 2010 at 9:07 AM

Hello,

If you've got some code based on Puma.NET or it's engine and would like to share it then it's possible to add a new capability to Puma.NET and make a new version. You may either provide me with the code or I may add you to the project as a contributor. Just send me a message if you decide and we'll agree on details.

Best regards,

Maxim.

share
Jun 10, 2010 at 9:53 PM

Hi,

Erik.  Thanks for offering the position, confidence, and charateristics code.  Is that code still available?

Thanks,

David

Jun 20, 2012 at 6:58 AM

I'm interested too.

Is there any update?

Thanks.

Jun 20, 2012 at 8:40 AM
Edited Jun 20, 2012 at 9:05 AM

Hello,

 

The code is still available for the retrieval of the word positions in the original puma library and can be obtained for a small fee.

However I spent nearly 6 full months to correct a lot of bugs in the original code of the library and I would recommend using the modifications I made if it is for a commercial use.

Best regards,

 

Erik

 

Aug 2, 2014 at 4:24 PM
Hello Erik Jan

How can I contact to you?

I'm interested obtain you library for word position feature.

Thank you,

Antonio
Jan 5, 2015 at 5:25 PM
Was this resolved ?

Can someone help?