How I can init the tesseract to work only with digits?

[ibrahimAlii]

Summary:
Well in my app I just want to initiate the tesseract to work only with digits, specifically arabic digits, Is there any way to avoid recognize characters and just recognize digits instead ?

I've use setVaraible but it's the same

baseAPI.init(dataPath, "ara");
        baseAPI.setVariable(TessBaseAPI.VAR_CHAR_WHITELIST, variable);
        baseAPI.setVariable(TessBaseAPI.VAR_CHAR_BLACKLIST, "ا ب ت ث ج ح خ د ذ ر ز س ش ص ض ط ظ ع غ ف ق ك ل م ن ه و ي ء هو !?@#$%&* >> << ()<>_-+=/:;'\\\"");
        baseAPI.setVariable("classify_bln_numeric_mode", "1");

I'm still getting characters in result.

[rmtheis]

Hmm, what's the value of variable in your example?

[ibrahimAlii]

@rmtheis It's all digits in arabic "٠١٢٣٤٥٦٧٨٩١٠"

[rmtheis]

Hmm, I would try setting just the whitelist value by itself, without setting the blacklist value or the other value.

If that gives you the same result maybe try asking on the Tesseract forum.

[ibrahimAlii]

@rmtheis Thanks, I still getting the same result, and I didn't get any useful response om Tesseract forum.

Any Idea?

[rmtheis]

1. Try different OcrEngineMode values.
2. Retrain or edit the training data. I don't know enough about this to be able to help.