Re: [tesseract-ocr] Trained font - always one letter wrong

Thanks for your effort! I tried language deu before and as you can see in your attached txts, there are some faults too. I could not eliminate them using freq- or user-words. But in general your result with deu is much better Than mine with v 3.05. -- You received this message because you are

Re: [tesseract-ocr] Trained font - always one letter wrong

Your image has text in German. You will get better results using language `deu` out of the box. Attached are OCR results using deu.traineddata from tessdata_best and tessdata_fast using tesseract-4.0.0-beta.1 run via command line. #tesseract sample.tif sample-deu-fast -l deu --tessdata-dir

Re: [tesseract-ocr] Trained font - always one letter wrong

Please provide a small sample image to test. ShreeDevi भजन - कीर्तन - आरती @ http://bhajans.ramparivar.com On Wed, May 2, 2018 at 11:26 AM, wrote: > Training doesn't work. If i use the characters "ä, ö, ü"

Re: [tesseract-ocr] Trained font - always one letter wrong

Training doesn't work. If i use the characters "ä, ö, ü" (which i need) in my training text, text2image says: WARNING: illegal UTF8 encountered and then creates an incorrect box/tif pair. This seems not to depend on my font, because with Arial it does the same thing. Can you help me to avoid

Re: [tesseract-ocr] Trained font - always one letter wrong

Use the latest version 4.0.0beta On Sun 29 Apr, 2018, 1:51 PM , wrote: > I did. Unfortunately they don't aswer... > Have you any advice for me, to improve the > training proccess? How many training texts should i use? Or is it possible > that there is a problem with

Re: [tesseract-ocr] Trained font - always one letter wrong

Check that your training text has enough samples for d. ShreeDevi भजन - कीर्तन - आरती @ http://bhajans.ramparivar.com On Sun, Apr 29, 2018 at 1:51 PM, wrote: > I did. Unfortunately they don't aswer... > Have

Re: [tesseract-ocr] Trained font - always one letter wrong

I did. Unfortunately they don't aswer... Have you any advice for me, to improve the training proccess? How many training texts should i use? Or is it possible that there is a problem with this font at all? Would help very much to find that out. Best regards Dave -- You received this message

Re: [tesseract-ocr] Trained font - always one letter wrong

Well, you should contact creator of traineddata . We have no clue what they did.. Zdenko 2018-04-25 14:55 GMT+02:00 : > Hello there, > > i don't know what to do anymore... > I want to use tesseract-ocr 3.05 for scanning documents, using the font > "Perfect DOS VGA 437

[tesseract-ocr] Trained font - always one letter wrong

Hello there, i don't know what to do anymore... I want to use tesseract-ocr 3.05 for scanning documents, using the font "Perfect DOS VGA 437 Win". Got a traineddata file for my font from trainyourtesseract.com, actual it works really nice but in every case the letter "d" isnt identified but