Inside Windows Platform | Inside Microsoft OCR Libraries

Play Inside Windows Platform | Inside Microsoft OCR Libraries
Sign in to queue

Description

In this week's episode of Inside Windows Platform, we talked with Ivan Stojiljkovic, the Dev Lead of the OCR team at Microsoft.

OCR is an extremely compelling developer scenario which can empower all sorts of amazing and useful mobile application. Microsoft's been developing OCR functionality for its applications for, literally, years. In order to get real world data, the OCR team first published their code as part of the Bing Translator app, which gave them an enormous amount of data, and compellingly delivers an experience of near real-time translation of any camera captured text.

Now, the OCR team is giving YOU the ability to leverage the power of character recognition to your applications. The OCR team has published their libraries to NuGet, for you.

Here are some links to the related materials:

 Let us know if you have any feedback!

@larryalieberman

Embed

Download

The Discussion

  • User profile image
    bcnz

    Disappointing this library isn't available for us WPF/Winform devs. Can't help but increasingly feel like a second-class citizens to Microsoft.

  • User profile image
    Nicolas​Dorier

    So disappointing, I heard "You can reference it with Nuget" I was so excited...
    Then I create a console app, and see the lib is only for mobile projects. (universal app, tablet and phone)

    Come on Microsoft... :(

  • User profile image
    MEK

    When this library was posted on the Windows 8 blog I tried to ask the same type of question as the ones above here but no answer.
    So I will try again:

    Why is this only available to Phone/Store apps?

    I assume that the OCR library is processing pixels in the form of bytes so this should be possible to do in any part of .Net?
    Is it because the library depends heavily on something in WinRT or is it to force us to only use WinRT?
    Or is the library already available in .Net under another name?

  • User profile image
    Flem100

    Thank you very much for this!

    I also mostly still create apps in WPF and therefore I use the ABBYYCloudOCRSDKService which is available on the Azure Marketplace. The free edition is for 100 pages per month I think.

    What are the differences between these 2 services?

    As far as I understand the basic differences are, that this is completely free and it supports multiple languages which I'm not sure that other service does. Also I prefere nuget packages over webservices.

    Are there some other major differences I should know about and is there gonna be a real desktop version?

  • User profile image
    PeterF
    @Flem100:
    I just came back from the ABBYY developer conference in Dublin. From what I can see the key difference is that the ABBYY cloud solution requires a data connection whereas the Microsoft package can be used offline. The ABBYY solution should have a key advantage in accuracy by using more CPU power, the ability to run more complex algorithms and better language support, for smaller tasks the Microsoft solution should be more than enough.
    For WPF and desktop apps there are many free open source options available.
    Interesting times in OCR space 😉

    Regards,
    Peter
  • User profile image
    Flem100

    Thank you so much Peter.

  • User profile image
    issam

    looks very usefull, so please consider WPF and desktop developpers too ,

    there are thousands and thousands of them :)

  • User profile image
    MichaelPolla

    Hi ! I'm also looking forward for a WPF compatibility :)

  • User profile image
    Tom

    Microsoft OCR is really excellent, I tried ABBYY and Tesseract but Microsoft OCR perform better for unclear documents. Please add it to .Net WPF and desktop.

Add Your 2 Cents