Abstract
Objective: Size constancy illusions are typically demonstrated using 2D stimuli in which one object is perceived as further than another despite being at the same distance. Our objective was to extend these findings using 3D reverse-perspective (RP), forced-perspective (FP), and flat stimuli (2D). Methods: In two experiments, observers viewed stimuli that had painted linear perspective cues that were either congruent (FP) or incongruent (RP, 2D) with the physical geometry. The RP stimulus was bistable (veridical or illusory), whereas FP and 2D only afforded one percept. Observers were asked to estimate the perceived size of a probe (disk or person) placed in one of two locations, by adjusting size of a second sample probe. In Experiment 1, observers were told that the stimulus could be bistable, were showed the two possible percepts, and were asked to adjust the probe when the percept was stabilized. In Experiment 2, new observers viewed the FP and RP stimuli, but were not told that RP could be bistable. Results and Discussion: In Experiment 1, person probes placed at the "more distant" position - as defined by the perspective cues - were perceived as larger for all stimuli, whereas disk probes were only perceived as larger at the "more distant" position for the 2D stimulus. Surprisingly, when perceiving the veridical shape of the RP stimulus, observers reported the "more distant" probe to be larger despite it being perceived to be closer, suggesting that observers continue to use the monocular perspective cues despite having binocular cues that provide reliable distance information. In Experiment 2, when observers were not told that the stimuli were bistable, observers again perceived the "more distant" test probe to be larger for FP and RP stimuli. In both experiments, observers were more accurate in judging the size of the disk than the human figure.
Meeting abstract presented at VSS 2014