Abstract
The aim of the present study was to determine the role of stereoscopic depth information in the orientation invariance of shape representations. Two experiments using sequential matching tasks were conducted. In both experiments, participants decided whether two stimuli shaped like bent paper clips were the same or different. The stimuli were rotated from 0 to 80 degrees around the vertical axis, and they could be presented without stereopsis, with normal stereopsis or with reversed stereopsis. The presentation mode was within-subject (n=12) in Exp. 1, and between-subjects in Exp. 2 (n=24). There were strong rotation effects in both experiments. In the first session of Exp. 1, the rotation effect was smaller with the stereoscopic presentation than with either the 2D or the reversed stereoscopic presentations. However, these effects were reversed in the second session of the experiment. In Exp. 2, the rotation effect was weaker with the stereoscopic presentation than with either the 2D or the reversed stereoscopic presentations in both sessions. These results indicate that stereoscopic depth cues may contribute to shape constancy. The loss of the benefit of valid stereoscopic information in the second half of Exp. 1 may be due to an adaptation to the inconsistency between the stereoscopic depth cues and the monocular depth cues in the 2D and reversed stereopsis conditions. This may have led participants to switch an initial strategy of trying to derive a 3D global representation of the shape to one of attempting to use 2D pictural cues to perform the task.
Meeting abstract presented at VSS 2012