Why pictures look right when viewed from the wrong place

Abstract

A picture viewed from its center of projection generates the same retinal image as the original scene, so the viewer perceives the scene correctly. When a picture is viewed from other locations, the retinal image specifies a different scene, but we normally do not notice the changes. We investigated the mechanism underlying this perceptual invariance by studying the perceived shapes of pictured objects viewed from various locations. We also manipulated information about the orientation of the picture surface. When binocular information for surface orientation was available, perceived shape was nearly invariant across a wide range of viewing angles. By varying the projection angle and the position of a stimulus in the picture, we found that invariance is achieved through an estimate of local surface orientation, not from geometric information in the picture. We present a model that explains invariance and other phenomena (such as perceived distortions in wide-angle pictures).

Toward a new theory of stereopsis.

“…focal length lens (Pirenne, 1970;Kubovy, 1986;Vishwanath et al, 2005). Moreover, we can all agree on whether a pictorial depiction of a 3-D object matches the real 3-D object-precisely what separates good drawings from not so good ones.…”

“…If they did, pictures and photographs would not serve the ubiquitous purpose they do. Moreover, there is extensive empirical evidence that pictures convey a consistent and measurable percept of quantitative depth and 3-D shape (Koenderink, van Doorn & Kappers, 1994;Koenderink, van Doorn & Wagemans, 2011;Kubovy, 1986;Vishwanath, Girshick & Banks, 2005). The crucial difference between pictures and real scenes appears to lie in the presence or absence of a qualitative attribute (stereopsis) regardless of the actual 3-D relations perceived.…”

“…Quantitative judgements of curvature-in-depth (Vishwanath & Hibbard, 2013), pictorial depth range (Wintjes & Pont, 2012), dihedral angle between planar surfaces (Cooper & Banks, 2012) and slant of planar surfaces (Erkelens, 2012) show little or no difference between binocular, monocular, and monocular aperture viewing. 31 Here we assume a picture with a "normal" projection and excluding situations where camera parameters themselves cause projective distortion (wide angle distortion, keystone effects, anamorphic effects; see Kubovy, 1986;Pirenne, 1970;Vishwanath et al, 2005) 32 An artist sometimes closes one eye when copying a real scene onto a picture, yet in the final picture the depicted objects appear to accurately match the 3D shape of the real object, even when viewed with both eyes. This lends more support to the idea that perceiving 3D surface shape is not linked to stereopsis.…”

Threefoldness

“…An odd fact about the psychology of picture perception is that if our position changes in front of the picture, our view of the depicted object does not change (Vishwanath et al 2005;Cutting 1987;Goldstein 1987;Halloran 1989;Pirenne 1970;Polanyi 1970;Wollheim 1980, pp. 215-216;Matthen 2005, pp.…”

A Study of the Different Uses of Colour Channels for Traffic Sign Recognition on Hierarchical Temporal Memory

“…Since it was not the purpose to test the learning and recognition capability of HTM, but to focus on how to deal with colour, a larger training set was not generated. These results do however indicate another difference between HTM and the human vision system, which is that the human vision system can learn when seeing items in one particular angle, and then recognise from a different angle [4].…”