Perceptions of Depth Elicited by Occluded and Shearing Motions of Random Dots

A computer-controlled display of random dots was used to study perceptions of depth. In this display, a field of stationary random dots surrounded a rectangular area in which random dots moved with uniform velocity in a single direction. The boundaries of this rectangle did not move. When dot motion was perpendicular to the longer boundary of the rectangle (occluded motion), the rectangle seemed to be behind the stationary background surround. Motion parallel to the longer boundary of the rectangle (shearing motion) made it appear in front of the surround. The relative lengths of the sides of the rectangle determined which effect predominated. Thus, for motion perpendicular to the long axis of the rectangle the occlusion predominated and naive subjects reported that the central area seemed farther away than the surround. For shearing motion parallel to the long axis, the subjects reported that the rectangle was closer than the surround and the strength of both effects also depended on the length-to-width ratio of the rectangle. If there was occluded motion along the long axis, as the length-to-width ratio increased so did the likelihood that subjects would report seeing the rectangle behind the surround. Conversely, with shearing motion along the long axis, increasing the length-to-width ratio increased the likelihood that the rectangle would appear unambiguously in front of the surround. Some subjects integrated the two cues with the resulting perception being a rotating cylinder. The occlusion effect was stronger than the shearing effect. In fact, this ‘far’ depth effect was so powerful that it tended to override conflicting depth cues such as height in the visual field or stereopsis. The ‘near’ depth effect produced by shearing motion was definite but these other depth cues could often override it.

Temporal Estimation in the Perception of Occluded Motion

Ss were shown a rectangular object which moved transversely across their field of view and passed behind an opaque screen. The purpose was to investigate some of the factors affecting estimates of the time required for the occluded moving object to travel a given distance behind the screen. The factors selected for study were (1) method of viewing the moving object (pursuit, static fixation), (2) background structure (homogeneous, textured), and (3) object size. According to previous studies, these variables affect the perceived speed of a moving object and might therefore be expected to affect estimates of the duration of occluded traversal. The results did not show statistically significant differences among experimental groups, although data trends are discussed. An additional finding was that Ss significantly overestimated the duration of occluded traversal, consistent with a tendency to overestimate traversal distance. The experiment is related to Michotte's studies of “amodal perception” and discussed in terms of Gibson's stimulus information approach to perception.