Abstract
Previous research has shown that perception of binocular 3-D object motion can be highly inaccurate (Harris, ECVP 2000, 2001). In these studies, observers were asked to maintain fixation on a stationary point. However, when viewing moving objects in real world situations, the moving object is often followed with the eyes. Are observers able to use information about their eye movements to help determine 3-D object motion?
In this study we compared a condition in which observers fixated on a stationary reference point whilst judging 3-D motion direction, with one where they followed the moving object with their eyes. The display always consisted of a pair of stereoscopically presented points (each subtending 8.3 min arc). Observers viewed trajectories ranging from 20 deg to the left of straight ahead, to 20 deg to the right. The perceived trajectory of the moving point was recorded by the observer, for each trial, by moving a pointer to reproduce the angle.
The average perceived trajectory angle for each observer, for each angle, was measured. Pooled results from 14 observers showed that very similar angles were perceived for each condition. Performance was highly inaccurate, as we found previously (e.g. angles of 20 deg are perceived, on average as 39 deg). We also tested whether the observer's precision was similar in the two conditions. We found that observers had a tendency to be more precise for the fixation condition rather than the eye movement condition.
The results suggest that there is no advantage to looking at either the fixation point or the moving point. The inaccuracies of trajectory perception shown in previous research cannot be improved on (or indeed are not impaired further) by the addition of eye movement information.