Detection of 3D curved trajectories: the role of binocular disparity

We examined the ability of observers to detect the 3D curvature of motion paths when binocular disparity and motion information were present. On each trial, two displays were observed through shutter-glasses. In one display, a sphere moved along a linear path in the horizontal and depth dimensions. In the other display, the sphere moved from the same starting position to the same ending position as in the linear path, but moved along an arc in depth. Observers were asked to indicate whether the first or second display simulated a curved trajectory. Adaptive staircases were used to derive the observers' thresholds of curvature detection. In the first experiment, two independent variables were manipulated: viewing condition (binocular vs. monocular) and type of curvature (concave vs. convex). In the second experiment, three independent variables were manipulated: viewing condition, type of curvature, and whether the motion direction was approaching or receding. In both experiments, detection thresholds were lower for binocular viewing conditions as compared to monocular viewing conditions. In addition, concave trajectories were easier to detect than convex trajectories. In the second experiment, the direction of motion did not significantly affect curvature detection. These results indicate the detection of curved motion paths from monocular information was improved when binocular information was present. The results also indicate the importance of the type of curvature, suggesting that the rate of change of disparity may be important in detecting curved trajectories.