The Role of Stereopsis, Motion Parallax, Perspective and Angle Polarity in Perceiving 3-D Shape

Abstract

We studied how stimulus attributes (angle polarity and perspective) and data-driven signals (motion parallax and binocular disparity) affect recovery of 3-D shape. We used physical stimuli, which consisted of two congruent trapezoids forming a dihedral angle. To study the effects of the stimulus attributes, we used 2 × 2 combinations of convex/concave angles and proper/reverse perspective cues. To study the effects of binocular disparity and motion parallax, we used 2 × 2 combinations of monocular/binocular viewing with moving/stationary observers. The task was to report the depth of the right vertical edge relative to a fixation point positioned at a different depth. In Experiment 1 observers also had the option of reporting that the right vertical edge and fixation point were at the same depth. However, in Experiment 2, observers were only given two response options: is the right vertical edge in front of/behind the fixation point? We found that across all stimulus configurations, perspective is a stronger cue than angle polarity in recovering 3-D shape; we also confirm the bias to perceive convex compared to concave angles. In terms of data-driven signals, binocular disparity recovered 3-D shape better than motion parallax. Interestingly, motion parallax improved performance for monocular viewing but not for binocular viewing.

Seeing more clearly through psychosis: Depth inversion illusions are normal in bipolar disorder but reduced in schizophrenia

“…2). Across all subjects, there were fewer DIIs when objects were viewed with stereoscopic cues than without ( F (1, 82) = 73.2, p < 0.001, ${\mathrm{\eta }}_{\mathrm{p}}^2=0.472$) and when objects were shown with realistic surface texture than without ( F (1, 82) = 63.6, p < 0.001, ${\mathrm{\eta }}_{\mathrm{p}}^2=0.437$), which is consistent with previously documented findings in healthy adults (Sherman et al, 2011). The effect of texture modulation was amplified with binocular viewing, F (1, 82) = 14.9, p < 0.001, ${\mathrm{\eta }}_{\mathrm{p}}^2=0.154$).…”

“…In the stereopsis viewing condition, subjects binocularly examined objects from a stationary chinrest. Stereoscopic viewing always occurred in the second half of the experiment because the visual system recovers depth better with stereopsis than motion parallax (Sherman et al, 2011) and because the aim was to minimize the chances that subjects would use their knowledge of a stimulus when making a response. The sequence of five objects was counterbalanced across participants and viewing conditions.…”

“…A classic and well-known variant of DII is the hollow mask illusion, in which a face eerily floats in front of its true position along the depth axis (Gregory, 1970). A host of factors contribute to DIIs including object familiarity, convexity-consistent texture and perspective cues, and a general bias to interpret any ambiguous 3-D surface as convex (Hill and Johnston, 2007; Ramachandran, 1995; Sherman et al, 2011). …”

Convexity Bias and Perspective Cues in the Reverse-Perspective Illusion

“…Mathematica 8.0.1.0 ( Wolfram Research, Inc., 2010 ) was used to record the duration that the observer pressed each of the two buttons and to compute the predominance of the veridical percept (percentage of the total viewing time in that percept). Predominance is a common measure of the strength of the veridical percept ( Dobias & Papathomas, 2013 ; Papathomas & Bono, 2004 ; Sherman et al., 2011 ). Data for each observer are the average of the two predominance values for each of the 3-minute trials for which the observer viewed each of the four conditions.…”

“…When viewing reverse-perspective stimuli ( Wade & Hughes, 1999 ), painted linear perspective cues can compete with bottom-up monocular (motion parallax, shading, lens accommodation) and binocular (disparity, vergence angle) depth cues, thus creating a bistable percept: The bottom-up cues favor the veridical depth arrangement, whereas the perspective cues elicit a depth-inversion illusion (DII), in which distant points on the stimulus appear to be closer than near points and vice versa ( Cook, Hayashi, Amemiya, & Suzuki, Leumann, 2002 ; Cook, Yutsudo, Fujimoto, & Murata, 2008 ; Dobias & Papathomas, 2013 , 2014 ; Hayashi, Umeda, & Cook, 2007 ; Papathomas, 2002 , 2007 ; Papathomas & Bono, 2004 ; Rogers & Gyani, 2010 ; Sherman, Papathomas, Jain, & Keane, 2011 ; Wagner, Ehrenstein, & Papathomas, 2008 ). Similar cases of DII can occur with either a hollow mask ( Gregory, 1970 , 1997 ; Hill & Bruce, 1993 , 1994 ; Hill & Johnston, 2007 ; Matthews, Hill, & Palmisano, 2011 ; Papathomas & Bono, 2004 ) or merely a hollow oval shape (“hollow-potato”; Hill & Bruce, 1994 ; Johnston, Hill, & Carman, 1992 ).…”