Localized text in images

[Parmeet]

Hi all,

I am working on recognizing text from device images. As i go through
papers, i find that it is always advisable to find text regions(text
localization) and then give it to any standard OCR engine to increase
the recognition accuracy.

I wonder, if Tesseract OCR itself try to localize the text first
before OCRing. If it is the case, can we somehow get the intermediate
result of localized text from tesseract OCR, so that i can compare it
with text localization results i obtained using other algorithms. This
way i want to make sure that the results of OCR i am getting are the
optimal one.

Also same is i wanted to ask for other intermediate results like
binarization, skew correction etc. Would there be a way to get all the
intermediate results before tesseract OCR actually work on recognizing
characters?

Thanks and Regards
Parmeet

[Vicky Budhiraja]

Hi Parmeet,

Sorry, I was not able to understand your question(s). Are you looking for a
method that can find text within graphics? Like text in Video frames or
Stylized text in a print-ad etc.?

Best Regards,
Vicky

--
Vicky Budhiraja
http://www.sitarasoft.com/

Hi all,

Thanks and Regards
Parmeet

--
You received this message because you are subscribed to the Google
Groups "tesseract-ocr" group.
To post to this group, send email to tesser...@googlegroups.com
To unsubscribe from this group, send email to
tesseract-oc...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/tesseract-ocr?hl=en

[Parmeet bhatia]

Hi Vicky,

Well, yes. I am actually trying to first localize text in images meaning finding text blocks within the image. I already put some algorithms in place which do the job for me. But i want to see if tesseract OCR also doing it before actually recognizing the characters. If it is the case, i want to compare my results with tesseract one. 

Also as text localization will vary from application to application and we may need to set different parameters in different scenarios. So i just thought it would be better if i optimize text localization for my application before giving those text blocks to tesseract OCR..

I hope it helps.

Thanks and Regards

Parmeet  

On Wed, May 25, 2011 at 10:26 PM, Vicky Budhiraja <vicky.b...@sitarasoft.com> wrote:

Hi Parmeet,

Sorry, I was not able to understand your question(s). Are you looking for a
method that can find text within graphics? Like text in Video frames or
Stylized text in a print-ad etc.?

Best Regards,
Vicky

--
Vicky Budhiraja
http://www.sitarasoft.com/

-----Original Message-----
From: tesser...@googlegroups.com [mailto:tesser...@googlegroups.com]
On Behalf Of Parmeet
Sent: Wednesday, May 25, 2011 16:51
To: tesseract-ocr
Subject: Localized text in images

--
You received this message because you are subscribed to the Google
Groups "tesseract-ocr" group.
To post to this group, send email to tesser...@googlegroups.com
To unsubscribe from this group, send email to
tesseract-oc...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/tesseract-ocr?hl=en

--
You received this message because you are subscribed to the Google
Groups "tesseract-ocr" group.
To post to this group, send email to tesser...@googlegroups.com
To unsubscribe from this group, send email to
tesseract-oc...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/tesseract-ocr?hl=en

--
Parmeet

https://sites.google.com/site/bhatiaparmeet/