The surfaces visible in the 3-D environment are basic elements in our perceptual representation of the physical world, but the projection of a surface onto the retina does not directly specify depth relationships within the surface. One might think of recovering a description of a surface as recovering the depth coordinates, but the perception of surfaces by human observers includes higher-order relationships, such as local orientation and curvature. The perception of these higher-order properties may be based on recovered depth coordinates or may be based directly on information in the stimulus. The proposed research examines the information used in the perception of relative depth, orientation and curvature for 3-D surfaces that are specified primarily by motion. Three series of experiments are proposed. In the first series the sensitivity of observers for detecting differences in depth, slant and curvature will be compared using a common task-judging which part of an object is different from the other parts in the attribute of interest. The second series of experiments will examine how judgments of depth, slant and curvature are affected by characteristics of the area surrounding the point at which the surface attribute is specified, including effects of discontinuities in the surface, of the total area available and of the area over which the attribute has a constant value. The third series of experiments will determine the effects of features surrounding the area in which the attribute is specified, including contour changes, occluding contours and the presence of a background. In each case the experiments are intended to be informative about judgments of each surface attribute separately and about how the effects of the variables under investigation might differ for different surface attributes. Overall, this research is intended to provide an understanding of how surfaces present in the 3-D environment, or in dynamic images simulating a 3-D environment, are represented perceptually by the human observer. This is important for the study of behaviors that depend on surface perception, including object recognition, object manipulation and navigation through the environment. This research will be conducted primarily in Bulgaria as an extension of NIH Grant No. R01 EY12437.
