This is kind of interesting as the University of Tokyo came out with something similar with their Tele-Existance/Crystal Vision/X-tal Vision projects.

(Remember the article on slashdot about the "cloaking" technology they were working on?)

Though their goal appeared to be to project a user's image onto a robot dummy and have the robot do things remotely. (Which is probably overkill anyway.)

A while back, I wrote an article on how this could be applied to education, interestingly enough, which can be read here:

http://tatewake.com/WhatIsTheFutureOfDistanceLearn ing.htm