Kinect for Windows gesture recognition

Question

I looked at the Kinect release notes and features for Windows , since I want to include gesture recognition in my project.

On the page above, the first line mentions that "Kinect for the Windows SDK allows developers to create applications that support gesture and voice recognition." The voice recognition API is available from the SDK and can be easily used. However, I don't think there are any gesture recognition APIs in the SDK. Skeleton tracking APIs should be used readily, but then they should be adapted for gesture recognition.

I saw a video from Windows Media Center beng controlled by gestures, etc. and other applications. I wonder if all these applications are custom-made and they have to write their own gesture recognition code.

I am currently using Codeplex's Kinect DTW Gesture Recognition from my project. I have two problems with this -> 1) It looks very fast, and when I turn on this application, my application throws an OutofMemory exception after some time (PC characteristics are quite high). 2) I can not say much about the reliability of the system. It works sometimes for some people, not for others.

I thought that if the APIs were embedded, it would be useful to switch to them. Are they available or what is permission?

Answer 1

I am doing it now for a school project. We had to create our own gesture recognition module. There is nothing in the API that will do this for you, but it will provide you with all the parts you will need to create features.

This article was a big help, http://blogs.msdn.com/b/mcsuksoldev/archive/2011/08/08/writing-a-gesture-service-with-the-kinect-for-windows-sdk.aspx . It talks about how to break gestures. If you have only a few gestures that you can hard code, this is trivial. We needed the ability to load and recognize user-defined gestures, but this article describes the basic structure that we used to do this.

Answer 2

Take a look here http://www.exceptontuesdays.com/gestures-with-microsoft-kinect-for-windows-sdk-v1-5/

Source code is available on this page.

It recognizes the following

switch (e.GestureType) { case GestureType.Menu: Debug.WriteLine("Menu"); Gesture = "Menu"; break; case GestureType.WaveRight: Debug.WriteLine("Wave Right"); Gesture = "Wave Right"; break; case GestureType.WaveLeft: Debug.WriteLine("Wave Left"); Gesture = "Wave Left"; break; case GestureType.JoinedHands: Debug.WriteLine("Joined Hands"); Gesture = "Joined Hands"; break; case GestureType.SwipeLeft: Debug.WriteLine("Swipe Left"); Gesture = "Swipe Left"; break; case GestureType.SwipeRight: Debug.WriteLine("Swipe Right"); Gesture = "Swipe Right"; break; case GestureType.ZoomIn: Debug.WriteLine("Zoom In"); Gesture = "Zoom In"; break; case GestureType.ZoomOut: Debug.WriteLine("Zoom Out"); Gesture = "Zoom Out"; break; default: break; 

Answer 3

I am also developing a gesture recognition application as a task at the university. After searching for good books, I came across these two, both written by Microsoft researchers:

(1) http://www.amazon.de/Programming-Kinect-Windows-Software-Development/dp/0735666814
(2) http://www.amazon.de/Kinect-Windows-SDK-Programming-Guide/dp/1849692386/ref=sr_1_10?s=books-intl-de&ie=UTF8&qid=1393944100&sr=1-10&keywords=kinect

They describe common methods to make gesture recognition possible with Kinect.

Answer 4

Try the FAAST API . This is an API where it can recognize body gestures and gives appropriate input interruption. Here you need to assign a specific key for a specific action. For example: suppose I use the "W" key to move forward, then I can assign this key to specific gestures using this API.