none
OCR using MODI

    Question

  • How to implement OCR using Modi ?

    I am facing the problem that the MODI is not soo accurate.

    i am using the code

    MODI.Document modidoc = new MODI.Document();
                        modidoc.Create(pictureBox1.ImageLocation);
                        modidoc.OCR(MODI.MiLANGUAGES.miLANG_ENGLISH, true, true); //selects the language
                        MODI.Image image = (MODI.Image)modidoc.Images[0];
                        rtxtbox.Text = image.Layout.Text;

    the code works fine,but still it is not accurate.

    Please help .

    Saturday, October 22, 2011 6:53 AM

Answers

  • MODI is quite accurate when you use it for images that have clear text and text in them is of a standard type but it fails to recognize characters that are not of the standard pattern try with the below image the results will be quite accurate


    --------------------------------------------------------

    Surender Singh Bhadauria

    My Blog

     

    Saturday, October 22, 2011 7:57 AM

All replies

  • MODI is quite accurate when you use it for images that have clear text and text in them is of a standard type but it fails to recognize characters that are not of the standard pattern try with the below image the results will be quite accurate


    --------------------------------------------------------

    Surender Singh Bhadauria

    My Blog

     

    Saturday, October 22, 2011 7:57 AM
  • Thanks a lot Surender Singh .You did a great job for me.
    MSDN Forum
    Monday, October 24, 2011 5:07 AM
  • The code and your use of MODI are correct from my perspective.  Plus you said it is working for you and you are getting results.  I am afraid this question is less technical and more technological.  Different OCR engines were designed with different algorithms and for different purposes, and will produce different results.  Sometimes comparing those engines is like comparing two different vehicles and their performance, and there is quite a big difference between a Honda Civic and Porsche Cayenne, even though they are both driveable vehicles.

    The quality of the OCR starts to be noticeable as documents become less perfect in quality.  For a crisp high quality document MODI and any other OCR can produce very high quality result and can be used very successfully.  For lower quality documents, specialized OCR engines have better more specialized algorithms and possibly specialized dictionaries to deal with restoration of damaged characters.  

    For example, feel free to try a free for developers yet commercial-level OCR Cloud 2.0 API available at www.ocr-it.com (http://www.ocr-it.com/ocr-cloud-2-0-api).  This may give you a different perspective at an alternative OCR and how dictionaries may help with recognition quality.

    Tuesday, November 01, 2011 5:08 PM