OCR: OCRTesseractDoOCR language

[ctn]

Hi,

Why this function "OCRTesseractDoOCR" always recognize the same text even if a different language has been selected for the same image?

Thank you.
Br, ctn

[ctn]

Could you please give me a feedback?
Thank you
Br, ctn

[Loïc]

Hi,

I've made test with success using different dictionaries.

Could you upload the image on which you have the same result ?

Best regards,

Loïc

[ctn]

Please find a "image" file (pdf) and the test result file (txt) attached.

Code snippet looks like:

Code: Select all

   ...
   TesseractDictionary ocrLanguage = TwainHandlerUtil.tesseractDictionary(language);

   if(Directory.Exists(@"ocrdata"))
   {
      string ocrText = _gdTwain.OCRTesseractDoOCR(ocrLanguage, @"ocrdata", "");
      _gdTwain.OCRTesseractClear();

      if(string.IsNullOrEmpty(ocrText))
      {
         throw new Exception();
      }
      return ocrText;
   }
   ...
   
   public static TesseractDictionary tesseractDictionary(OcrLanguageEnum language)
   {
      TesseractDictionary ocrLanguage;
      switch(language)
      {
         case OcrLanguageEnum.ENGLISH:
            ocrLanguage = TesseractDictionary.TesseractDictionaryEnglish;
            break;
         case OcrLanguageEnum.FRENCH:
            ocrLanguage = TesseractDictionary.TesseractDictionaryFrench;
            break;
         case OcrLanguageEnum.GERMAN:
            ocrLanguage = TesseractDictionary.TesseractDictionaryGerman;
            break;
         case OcrLanguageEnum.ITALIAN:
            ocrLanguage = TesseractDictionary.TesseractDictionaryItalian;
            break;
         default:
            ocrLanguage = TesseractDictionary.TesseractDictionaryGerman;
            break;
      }
      return ocrLanguage;
   }

   public enum OcrLanguageEnum
   {
      GERMAN,
      FRENCH,
      ITALIAN,
      ENGLISH
   }


Thank you.
Br, ctn

[ctn]

Attach files again.
Br, ctn

Seems that "Upload attachment" does not work.
Br, ctn

[Loïc]

Hi,

Send me your images at esupport (at) gdpicture (dot) com.

Best regards,

Loïc

[ctn]

Hi Loïc

An email has been sent.
Thanks
Br, ctn

[Loïc]

Hi,

I got your mail, thanks !


For me the OCR result are quite different regarding the dictionary.
There is no problem.

IE in first line using first German dictionary then English dictionary :


German

Die Rahmenprogramme für Forschung und technologische Entwicklung (FRP) sind das

English

Die Rahmenprogramme fur Forschung und technologische Entwicklung (FRP) sind das

Kind regards,

Loïc

[ctn]

Hi

I thought that only recognized words in the expected language will be returned.

Br, ctn