Abstract
Conventionally, Figure-Ground Organization (FGO) has been studied by asking the subject to decide whether a 2D region in a picture looks more like a figure or ground. The main assumption behind these studies was that, from a computational point of view, FGO is a 2D operation. Two years ago we reported that performance in a FGO task was somewhat higher when the figure was perceived as a symmetric shape having volume. This was done by asking a subject to recognize the 2D shape of a figure or that of ground. In the present study we directly tested the role of three-dimensionality of a symmetric shape in FGO, by using a shape constancy task. If the observer perceives a given region in a stimulus as a figure (object), then she should be able to recognize the shape of the object. At the same time, we don’t expect the observer to be able to recognize the shape of the background. In order to make sure that the subject is recognizing a 3D shape of an object, rather than a 2D shape of a region on the retina, we used shape constancy task. The observer memorized 3D shapes of two test stimuli in the first presentation and judged whether either of them was identical with the 3D shape of a stimulus in the second presentation. In "same" trials, the 3D shape was rotated around the vertical axis by a 20 deg angle. The stimuli were viewed binocularly. In one condition, the stimulus was perceived as a 3D symmetric "polyhedron". In the other condition, the stimulus was perceived as a 3D curve surrounding "empty space" between two polyhedral objects. Four subjects were tested. The results show that the performance with "polyhedron" is much better (d'=1.96) than with "empty space" (d'=0.79).
Meeting abstract presented at VSS 2012