No, drivers for the image-processing-chip in the Kinect have been open-sourced, which give you a depth-map of the image. That's only a small part of the work, you'll still need to do the (absolutely non-trivial) work of tracking objects, and determining body-parts and stuff, which needs to be close to perfect, or it'll be an excercise in frustration. MS got the Kinect's hardware somewhere in 2006-7, and spent years making it actually gameplay-worthy...
As for Windows-on-Arm: I expect the collection of software available for it to be very, very limited for quite a long while, basically being limited to MS-stuff, open-source stuff (which you'll probably have to compile yourself, because no-one else has), java, pure .net-apps, and (probably) WP-silverlight-apps, but the amount of worthwile java and .net-apps is pretty limited, and most stuff doesn't have a touch-UI. Commercial developers will barely look at it before there are a significant amount of devices out there, most open-source developers won't even consider buying one, and compiling open-source stuff on windows isn't particularly easy, so won't happen all that much...
There might be some use for it in corporate environments, with mostly custom, in-house software, but that market is tiny. (just take a look at how "well" the Windows-tablets, which are mainly targeted at that market, have managed to sell for the last 10 years..., or PDA's for that sake)
Oh, and MS is going to get serious driver-issues, as linux is by now pretty much the sole OS for most high-end ARM-chips, and most chip-manufacturers won't bother writing good drivers for Windows-on-Arm, and nobody will be able to fix the bugs...
Furthermore, as always when MS announces something, it won't be released within the next few years, and when it finally does get released, it's years behind the rest of the market, and nothing like the original announcement. I reaally don't get why people keep falling for that trick time and time again...