Foreshortening leads to overestimation of angles
Marta Wnuczko · TSpace (University of Toronto) · 2015
The hypothesis tested here is that vision's misuse of foreshortening, such as in an optic array's texture gradient, may lead to errors in the perception of angles. While errors may increase with optical slant, they may also be reduced by the use of optic information for the shape of surfaces. The first hypothesis was tested in five experiments. In the first four, observers estimated acute angles between a line receding away from the observer into depth - the z dimension -- and a line oblique to the z-dimension presented on the ground (Experiments 1 and 2) or a wall (Experiments 3 and 4). In Experiment 5, observers estimated angles between a slanted edge of a square platform and an oblique line lying on top of the surface. Foreshortening was increased by increasing the observer-to-target distance along the z-line (Experiments 1-4) and by reducing the distance from the eye to the surface bearing the target (Experiments 2, 4, 5). As expected, target angles were increasingly overestimated with foreshortening in Experiments 1-4. In Experiment 5, target angles on the slanted surface were not overestimated (they were underestimated modestly). Further, geographical slants were overestimated for surfaces. This was true for surfaces covered in different textures (randomly or cardinally oriented texture elements), in line with the foreshortening hypothesis.