Motion based AR using Kinect

Hi,
I would like to get your valuable recommendations for achieving something like this :

I would like make something like this but on a smaller scale , with 42+ inch screen, for kids shop.
From my research i found that kinect,asus xtion openni sensors can be used for interactions like this.
Can we achieve this kind of results using kinect for windows ?
Would i need a seperate webcam to give better video stream or kinect has inbuilt?
I am planning to simplify by using assetstore plugins supporting openni or go for commercial zigfu zdk.

I am not very experienced in these kind of AR. So please give you opitions,recommendation and adivices.
Thanks in advance.