[tesseract-ocr] Re: Passport MRZ characters OCR

[Tamás Bosnyák]

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

Thanks

2014. augusztus 8., péntek 10:04:42 UTC+2 időpontban mail2kul a következőt

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

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

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

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?

[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.

Post by 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.