This sounds like (the reverse of?) something I saw recently, "Projection Mapping on a Moving Object," demonstrated on a bouncing ball:
http://www.youtube.com/watch?v=ZuSUHuSceYc
The tech's called Lumipen, and they use something called Saccade Mirrors to do it.
Given perfect eye tracking, that could work for the vestibulo-ocular reflex case where the eye counter-rotates to keep the world stable during head rotation, also given that no object is close enough for the parallax from the eye moving in space during head rotstion to be noticable.
It would not fix smooth persuit tracking of a moving object while the head is stationary though, as you could not selectively shift the moving object to match eye rotation while also keeping the background stationary.
Mirrors were my first thought as well, though in a slightly different way; you could have the mirror track the scene instead, counteracting head movement during a single frame's persistence and then snapping back between frames to continue the motion (probably with some display that could provide partial-frame persistence). It seems like this would address VOR, which from my reading of the article sounded like a bigger problem than saccading. A quick googling turned up piezo mirrors [1] with specs that look like they'd be in the right ballpark.
[1] http://www.physikinstrumente.com/en/products/prdetail.php?so...
Aside, this stuff is really fascinating, and it's great to see the field of human-computer interaction pushing up against formerly unknown biological phenomena.