All we need now is a way to directly project images onto our retinas for total immersion.
... and to detect the shape of your lens (in your eye) to then update the project depth-of-field so that you can focus on stuff, thereby avoiding the 'we are projecting an image of a screen in front of your face' problem, which is a major hinderance to the immersive experience. Do this in less than about 12 milliseconds? Hmm.
I'm guessing there are (or shortly will be) a number of patents covering possible methods of detecting the current distortion of your eye's lens.