The display thingie is a combination of things that do exist, although such combination is a bit of a stretch.
It is capable of detecting eyes in the room. For comparison, most digital cameras, including those on smartphones, recognize faces. Red-eye removing software is often capable of detecting eyes. Of course, all of these are not really reliable enough to be trusted with the outcome of the secret mission of breaking in Kremlin, but that doesn't mean that the technologies cannot be improved.
The next thing it does is display an adapted version of an image, so that it is always a 2D projection adapted to the position of the eyes of the viewer. This part is a bit thin, because normal 3D vision is achieved by eyes receiving two slightly different pictures, but given the simplicity of the scene (an empty hallway), it might be feasible. I'd say it would be quite convincing if the guard wasn't, at moments, staring really intently at the projection, but this kind of exaggeration is the most common element of movie suspense, so we can forgive that.
It also renders each scene very quickly, as the guard moves. Not really a stretch, since common PCs do that in many modern (and even not so modern) games. This is even simpler, due to the aformentioned simplicity of the scene.
So, all you need is a perfectly silent mechanism to setup the projection screen, accurate 3D imaging of the environment to prepare the scene (I believe those also exist, although I suspect today's technology is not capable of ignoring two agent's standing in the way of some of the nearer parts of the wall ;-)), and plenty of luck, so that the guard doesn't notice that the scene is actually 2D. You also need a perfect layout of Kremlin, so that you can prepare the screen that will fit the hallway width and height perfectly, and similar minor details which, in my opinion, would be much harder to fix (in real world) than the technology itself.
Funny, you didn't mention a device which they used in that scene to send the sound (of water dripping) to a distant corner. I have no idea if that exists, nor if it is possible.
The lenses need extreme miniaturization (so that a computer would fit there, without being in the way of the one using such lenses), some way to record the image (this one is beyond today's technology, since we still rely on lenses which need depth between them), some WiFi device, etc. I'm not sure if "a thin layer of glass capable of recording stuff" is physically possible, but I believe the rest is (it may even exist, but it's not on the market).
The lenses remind a bit of a miniaturized Google Glass, but notice that even Google Glass has a "fat" device (in front of the right eye):
Also, human eye "records" the images by using lenses, so it also needs depth. This is why I find this part the least convincing one.