It sounds like a really interesting idea, but I don't have any personal experience with it or know of any related resources -- I'd love to know more myself if you're able to find anything or if you end up trying it out though!So nobody commented on the possibility of using Kinect to generate animated 3D models. I know it can be done because at the latest Game Development Association meeting in my area another developer actually brought a Kinect and demonstrated how to capture animations with it and then generate animated 3D models for use in Unity. The only thing I am wondering is, how feasible this really is, and how much work would it involve? Would it be more difficult than, say, modifying an existing stock asset?
I've seen it used professionally at my last job. There's plugins for animation software such as MotionBuilder that can directly interface with the Kinnect as a motion-capture device. AFAIK, there's also some open-source standalone apps to do the same.
It's the same as any motion-capture system. Raw mo-cap data is never fed directly into a game, ready to use. Mo-cap data is a starting point for an animator, who then has to do a lot of work with it in their animation package (MotionBuilder, etc) to clean it up and get it ready for use as a final game-ready animation.
N.B. that even professional-grade mo-cap systems -- where you're spending tens of thousands on rent, actors and operators -- you still need a skilled animator to take the recorded data and turn it into a game-ready animation. Kinnect is a poor-man's replacement, so the data is worse quality, so there's even more work for the animator to do in the "clean-up" phase...
It's a great time-saver for an animator, but it's not a replacement for having a skilled animator.