Re: Passport MRZ characters OCR

[Sven Pedersen]

You should provide a sample image (or due to the sensitivity of the subject, slices of images showing those characters). That will help us to see what your input looks like. Scanning at a good resolution should yield decent results -- see the FAQ.

--Sven

On Fri, Jan 11, 2013 at 3:15 AM, sav <savan....@gmail.com> wrote:

Dear All,

    I am now needing to OCR the MRZ characters on the Passport. These characters are in mostly OCR-B font. 

    I use two url as a reference : 

    1. http://code.google.com/p/tesseract-ocr/wiki/TrainingTesseract3

    2. http://michaeljaylissner.com/blog/adding-new-fonts-to-tesseract-3-ocr-engine

    Now the problem is that box file is displays all true characters but when I try to ocr that passport or any other document which has same font then it was not recognize the all true characters.

    Mainly it gives wrong output for O,0,W,M,Z,2,4,V characters.
    Can anybody give me some advice on this, or image pre-processing technique to improve the OCR result? Thank you all!

--
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

--
``All that is gold does not glitter,
  not all those who wander are lost;
the old that is strong does not wither,
  deep roots are not reached by the frost.
From the ashes a fire shall be woken,
  a light from the shadows shall spring;
renewed shall be blade that was broken,
  the crownless again shall be king.”

[Rupam Bhattacharya]

You have reached a very decent success. I am still strugging at much lower accuracy. Can you please share the training file?

On Tuesday, 22 January 2013 13:56:26 UTC+4, sav wrote:

Thank you for reply. 

I got 95-96% accuracy in few passport, and 100% for remains to detect MRZ.

And size and resolution of that image is 1900-2000 x 250-300 pixel and 150-300 dpi.

so, now please tell me what to do? 

[Rafael Benitez]

I too am trying to read the passport MRZ.

could you share the tessdata/ ?

[Surya Rajput]

Hi,

Would be appreciate if you could you please share training data for OCR-B font.

I am struggling  it with , done 4 round of tesseract training but no success.

Thanks

[Surya Rajput]

I am actually looking for training data for MRZ font of passport and other travel document.

[mail2kul]

My accuracy is 80-90% on iOS. can anyone share me tips to get 100% result for passport MRZ reading

[Tamás Bosnyák]

I know it is an old question, but do you have any good results with MRZ traineddata? 

Thanks

[shilpa rane]

hello,

          I am using Tess4j for tesseract-OCR. I have traineddata for passport's mrz. While I am trying to read mrz portion of image am getting spaces in the output. For example, In image contains "CLIFFORD" and I am getting "CLI F FORD" as a output. Please help me to solve this problem.Can  someone explain me how tesseract adds this spaces in the output.Thanks in advance. 

[Anton Prabhakar]

Hello Shilpa,

Could you kindly share the traineddata for Passport MRZ ? It would be a great help.

[buffey tt]

could you please help to share the training data of passports?