Can I improve result of the TextRecognize by training the machine which is behind? Id like to work with a lot of scanned texts from several (real) typewriters. It seems to me that TextRecognize gives much better results for computer fonts than for real typewriter fonts.

  • 1
    $\begingroup$ have an example? $\endgroup$ – george2079 Mar 9 '18 at 20:15
  • $\begingroup$ You could probably train your own network, as in 143691. $\endgroup$ – anderstood Mar 9 '18 at 20:15
  • $\begingroup$ Anderstood - Id like using the TextRecognize features but to train the machine behind (Tesseract) to my specific data. I was not meaning creating my own neural net from scratch $\endgroup$ – CJoe Mar 9 '18 at 21:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.