The role of linear perspective cues on the perceived depth magnitude of the surfaces they are painted on: proper-, reverse-, and flat-perspective paintings.

Abstract

Evidence from several studies suggests that perspective cues influence the perceived three-dimensional (3-D) layout of the surface they are painted on. The purpose of this study was to test how perspective cues influence the perceived depth magnitude. We rendered the same linear-perspective scene on three different surfaces: proper perspective, with a 3-D structure that was congruent with the painted scene; flat perspective, incongruent with the scene; reverse perspective, opposite to the scene, producing two competing stable percepts (veridical and illusory). We varied binocular disparity by using three different sizes for each type of stimulus. Observers assessed the magnitude of the perceived depth within each of these stimuli. Accuracy improved with increasing stimulus size that covaries with binocular-disparity magnitude. Generally, the magnitude of the perceived depth of stimuli painted with perspective cues was larger than the physical depth of the stimulus regardless of stimulus type (proper, reverse, flat). Further, depth magnitude tended to be larger when depth cues were congruent (proper) as compared with opposite (reverse) or incongruent (flat). There was no difference in perceived depth under the different percepts (veridical and illusory) for the reverse-perspective stimulus, suggesting that depth is assessed by the stimulus structure rather than by the percept obtained.