Help

[Ishak DÖLEK]

I want to train the manuscripts using tesseract v4. 

Do I need to convert the image of manuscript  into binary pictures to train? What do you suggest to convert binary pictures if you need to?

Thank in advance.

Sincerely

[Zdenko Podobny]

Please provide example image you try to OCR.

Zdenko

št 31. 10. 2019 o 22:34 Ishak DÖLEK <ishakd...@gmail.com> napísal(a):

--
You received this message because you are subscribed to the Google Groups "tesseract-ocr" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tesseract-oc...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/CAA%3DdkuZA0MPw%2BDS6UyPt4rp4LeDBm%2B27J2MKXf37kjvOjjpmxQ%40mail.gmail.com.

[Ishak DÖLEK]

I'm sending the image below. All image are like this.

Zdenko Podobny <zde...@gmail.com>, 1 Kas 2019 Cum, 11:04 tarihinde şunu yazdı:

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/CAJbzG8yLjCbbJLQnPxiCbRr8y1PBQiyjmKGwy3Z3Hu87%2BzM5YQ%40mail.gmail.com.

[Zdenko Podobny]

No, you do not need binarize image if you plan train tesseract from custom images.

Zdenko

pi 1. 11. 2019 o 9:59 Ishak DÖLEK <ishakd...@gmail.com> napísal(a):

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/CAA%3DdkuZQpWTwipNEZiQx_e-g8Xo%3DfnNRrbe677YUrxpqGY9-ZA%40mail.gmail.com.

[Ishak DÖLEK]

I would like to create a traineddata to recognize the Ottoman manuscripts.
I have prepared both synthetic and orijinal (tif/box) data for this purpose. I also prepared a dictionary list.
For this, it would be correct to convert the Ottoman mannuscripts images into binary images?
or what would you suggest me to do this task?

Thanks in advance.

Zdenko Podobny <zde...@gmail.com>, 1 Kas 2019 Cum, 13:24 tarihinde şunu yazdı:

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/CAJbzG8z94PZ_7LybHE_YHRshVEgYk7ou4eXzW4Mpsin%2BqP-y_w%40mail.gmail.com.