Read Text with OCR - error in Reading Korean

[HongJooChoi]

Dear Community

Dose the "Read Text with OCR" action support only English? or
Should we do something to give it an ability to read non-English, that is, Korean in my case?

To use the "Read Text with OCR" action is the last resort for me as the "Recognize Text" action didn't work at all, nor could I find any relevant font from font searching.

While I was testing the Read Text with OCR action from Read Stage,


I came across the following error:


I did check out the corresponding folder and found there's seems no dataset or relevant data for Korean.



Your help would be much appreciated.


Best regards

------------------------------
HongJoo Choi
------------------------------

[JuanSalazar]

I'm not sure where you can find the .user-patterns and .user-words  for Korean, but the .traineddata can be found in this repository https://github.com/tesseract-ocr/tessdata
hope it's helpfull!

------------------------------
Juan Salazar
------------------------------

[ShirleyZhang · Accepted solution]

Yes, Read with OCR could be made compatible with Korean. It's supported by tesseract.

You just need to download the Korean Traineddata from here: Traineddata Files for Version 4.00 + | tessdoc (tesseract-ocr.github.io)

And add the file to here: Optical character recognition (OCR) (blueprism.com)

Let us know if it works. Good luck!

------------------------------
Shirley Zhang
------------------------------

[HongJooChoi]

Dear @JuanSalazar

I confirm that it works.
Thank you for your help.


Best regards​

------------------------------
HongJoo Choi
------------------------------

[HongJooChoi]

Hi, Zhang

Thank you for your help and also for the reference to the portal which I didn't know.
It works !


Best regards

------------------------------
HongJoo Choi
------------------------------

[BohyonHwang]

I also found kor_vert.traineddata which supports vertical writings. Awesome!

------------------------------
Bohyon Hwang
Director
블루프리즘
Seoul
+82262033950
------------------------------