word review

[Thilanka Kaushalya]

Hi,

       I'm using Tesseract for my letter recognition project and currently the recognitions is quite good.
The letters are hand written.But there are some problems when I used it to recognise the letter "O" and
number "0". These letters are used in data areas as the fields that enter names. So names cannot have any
numbers with it. And when we are using the the system of the data fields as date of birth it only contains
numbers. So I'm willing to give restriction to the recognition system saying that the corresponding data fields
have only numbers or the letters.
       And also I'm willing to review the recognised letters with the possible words so we can improve the accuracy
of the data. But I don't have any idea about how to do that.  

Can some one help me. Thank you.

Regards,
Thilanka.
--
http://coders-view.blogspot.com/
http://thilankagekawuluwa.blogspot.com/
http://twitter.com/thilanka_k

[Joe K]

Hey Thilanka,

I ran into a similar problem when I only needed it to look at
hexidecimal values. What I ended up doing was creating a separate
"langauge" that only contained the specified characters. So you could
create a langauge of numbers and a language with letters and use
tesseract to read each part of your image using the appropriate
language.

The web address below shows you how to train tesseract for a specific
language. Hope this helps.

http://code.google.com/p/tesseract-ocr/wiki/TrainingTesseract

[Moffette]

Hi,

An easier way to deal with number only or letter, is to use this from
FAQ (http://code.google.com/p/tesseract-ocr/wiki/FAQ):
----------------------------------------------------------------------------------------------------------------------------
How do I recognize only digits?

In 2.03 and above:

Use

TessBaseAPI::SetVariable("tessedit_char_whitelist", "0123456789");

BEFORE calling an Init function or put this in a text file called
tessdata/configs/digits:

tessedit_char_whitelist 0123456789

and then your command line becomes:

tesseract image.tif outputbase nobatch digits

Warning: Until the old and new config variables get merged, you must
have the nobatch parameter too.
----------------------------------------------------------------------------------------------------------------------------

For the second part : " I'm willing to review the recognised letters

with the
possible words so we can improve the accuracy "

If you are using a 2.0X version you could use the eng.user-words (a
user dictionary) as it's suggested in the FAQ (http://code.google.com/
p/tesseract-ocr/wiki/FAQ)

----------------------------------------------------------------------------------------------------------------------------
How do I provide my own dictionary?

Easy: Replace tessdata/eng.user-words with your own word list, in the
same format - UTF8 text, one word per line.

More difficult, but better for a large dictionary: Replace tessdata/
eng.word-dawg with one created from your own word list, using
wordlist2dawg. See the TrainingTesseract wiki page for details.
----------------------------------------------------------------------------------------------------------------------------

> > --http://coders-view.blogspot.com/http://thilankagekawuluwa.blogspot.co...

[Thilanka Kaushalya]

Hi Joe and Moffette,

             Thanks for the tips you provided. those are very helpful for me. These days

I'm testing your instructions. Thanks again. 

regards thilanka

 

 Topic: word review

Joe K <joekar...@gmail.com> Mar 08 11:02AM -0800 ^

 
Hey Thilanka,
 
I ran into a similar problem when I only needed it to look at
hexidecimal values. What I ended up doing was creating a separate
"langauge" that only contained the specified characters. So you could
create a langauge of numbers and a language with letters and use
tesseract to read each part of your image using the appropriate
language.
 
The web address below shows you how to train tesseract for a specific
language. Hope this helps.
 
http://code.google.com/p/tesseract-ocr/wiki/TrainingTesseract
 
 
 


 

Moffette <omof...@gmail.com> Mar 08 12:26PM -0800 ^

[Thilanka Kaushalya]

Hi Joe and Moffette,


         I'm recognising the data from a hand written form, and the scenario is extracting the
letters one by one and sending the each letter to the tesseract seperately. So the recognition
is done letter-vice. So I can,t use the dictionary file for the word reviewing in that case.
****************

How do I provide my own dictionary?
 
Easy: Replace tessdata/eng.user-words with your own word list, in the
same format - UTF8 text, one word per line.
 
More difficult, but better for a large dictionary: Replace tessdata/
eng.word-dawg with one created from your own word list, using
wordlist2dawg. See the TrainingTesseract wiki page for details.

***********************
          Is the Tesseract output the words only included in the above mentioned libraries.
If so can I send the set of recognised letter again to the Tesseract as an image to review
it to the defined domain of per-defined words.

          Or else can you give some instructions about a method to do how.

Thanks and regards,
Thilanka.