3D Selection Vault

VOR Depth Selection

Also known as Gaze & head-shake, VOR depth estimation

Look at the target and shake your head: the ratio of eye to head rotation during the shake reveals how far away you are looking, which tells apart objects that overlap along the gaze line.

How it works

While the user keeps gaze on a target and deliberately shakes the head, the vestibulo-ocular reflex counter-rotates the eyes; the eye-to-head angular velocity ratio (VOR gain) is larger for nearer fixation points, so it gives the gaze depth after a per-user depth calibration. It works with one eye and without accurate gaze calibration (pupil movement in the eye image suffices). The head-shake doubles as the selection trigger, and the candidate whose depth matches the estimate is selected. In a VR study (N=11, HTC Vive with Tobii eye tracking) with pairs of partly overlapping targets at 0.3–7 m, selections took 0.56 s on average; accuracy matched vergence-based depth up to 1 m and was 18% better at larger distances. Natural head shifts without a deliberate shake did not work.

Papers