Call me skeptical, but this is old hat and isn't even very good. Watch the guy's hand. It always is obscured when he moved it "behind but in front" of the object.
sgi O2s and Amiga A4000s have been doing this for years. Watch hockey on Fox and see how they track the puck. Or even football with the yellow first-down markers. Notice how the players never move under the enhancements. They are always on top of them.
This video has layering problems because of that. It looks like whomever set it up didn't know what a layer was and just pasted the object right over the top of the film.
My sgi O2 can do that on 3 simultaneous video streams and it's circa 1997.
It is different in that it "tries" to do 3D but it doesn't do it very well. All of this "front-end" viewer work likely took hours and hours of backend work behind the scenes to even make it look good.
Like I said, maybe it's the skeptic in me. I've seen literally hundreds of presentations doing something "new" in tech and they normally all have the slick front-end and horrendous back-end problem in common. Almost without fail.
Here's a direct link for those that don't want to use their browser. (Right-click, Save as)
Yes, I know I'm oversimplifying what's going on here, but ultimately it isn't that impressive to me.