[Summary Week 6] The importance of stereo and eye-coupled perspective for eye-hand coordination in Fishtank VR

In their paper, Arsenault and Ware discuss the relative importance of stereo-based and motion-parallax based depth cues in a virtual environment. Stereo depth cues are provided due to disparities between the pair of images formed in the left and the right eye, for the same scene. Motion-parallax also provides a monocular depth cue, arising due to the difference in the relative motion of a close by object and a distant object.

First, the author’s highlight the robustness of the human perceptual system that allows for correction of perspective distortion. They also mention the importance of modeling tactile feedback in a study comparing efficacy of various depth cues, stating, that such feedback could play an important role in recalibration of the human depth-perception system.

For judging the relative importance of the two depth cues, the authors propose a cylinder tapping task, which is a slight modification of the classic Fitts’ tapping task (that led to establishment of Fitts’ law and the establishment of a metric for such tasks – the index of difficulty). Stating the technical difficulties that prevent the realization of a high field of view stereo experience, the authors describe their experimental setup that is essentially a virtual environment with coincident visual and haptic display. Using a PHANToM system for providing a haptic workspace, the authors use LCD shutter glasses and tracking technologies to provide stereo and eye-position aligned perspective vision respectively.

The authors describe their experiment which consists of tapping a sequence of target cylinders, starting from a home cylinder. The target sequence had cylinders placed and sized so as to yield 4 index of difficulty values. A trial sequence set (a set of 4 sequences with different combinations of Head Tacking(HT) and Stereo(S) enabled/disabled) was followed by 4 actual sequence sets.

Carrying out this experiment on 19 participants, the authors state that they found that availability of stereo viewing provided a greater boost to tapping speed, as opposed to the availability of head-tracking (and thus motion-parallax). They also found evidence of both of these cues having increasing advantage for increasing difficulty of the task. This leads the authors to surmise that for visually guided reaching tasks, stereoscopic viewing furnishes more vital depth cues as compared to head-tracking.