Since Microsoft released the official Kinect SDK for Windows I have been doing a tiny bit of mucking around with it.
I am working on a poor-man’s “Green Screen” app which uses the Kinect’s skeleton tracking to mask out the background. I got the first stage of it working today.
I’ll post more videos as the software progresses.
guysherman.com 
Interesting – what made you decide to do this using skeleton tracking? The official SDK already includes player segmentation data in the feed. I know that the big problem with this is that it can’t be applied directly to the image frames but I would have thought that some kind of hack could be implemented until the SDK supports the translation from depth map to image map natively.
You may also be aware that this can be achieved quite easily using OpenNI and the hacked Primesense driver for the Kinect – since those libraries provide the segmentation data and do the mapping for you, it’s an easy task to get the “green screen” effect. The resulting image is pretty low res though, as the open source libraries only support image capture at up to 640×480 – I’m hopeful that something using the official SDK will give a better result.
Hi Jonathan,
It is using the player segmentation data, it’s just that you have to turn on Skeleton Tracking to get it – I’m not actually using the skeleton joints to generate the mask. My next strategy for the mask will be a screen-space edge detection on the depth image to pick out objects that stand out from the background. Haven’t really thought about how I’ll get it to determine wether an object is human or not, but I think some sort of centroid tracking could be used to follow objects once they are detected. I’m thinking of putting it up on codeplex or GitHub in the near future, but I’m not quite there yet.