Item

Abstract

It has been known for a long time that many cues contribute to the perception of 3D shape
from 2D images, such as shape from shading, textures, occlusions or reflection of the surrounding
environment. However, little is known about the influence of lighting conditions on
the correct mental reconstruction of 3D shapes. In order to investigate this, we have run a
set of experiments asking participants to report differences in surface orientation of unknown,
smooth surfaces, using different methods. The first experiment consisted of a 2AFC in which
subjects had to identify which of two test objects had the same shape as the target. The stimuli
were computer generated irregularly-shaped smooth surfaces, illuminated by a single point
light source. For both test stimuli, the position of the light sources could either be different
from or the same as the target. Results show that, as the amount of shape difference became
smaller, participants were more and more biased towards choosing the match shape lit by the
same source as the target. In the second experiment, participants had to report the perceived
orientation of the surfaces at various locations by adjusting gauge figures.. The surfaces could
either be the same or slightly different and the light source of each shape could either be the
same or offset by 90 degrees horizontally. Participants’ matches revealed large differences in
perceived surface orientations when the lighting was different, even when the shapes were the
same, confirming the first results. Our findings show that lighting conditions can play a substantial
role in the perception of 3D structure of objects from their 2D representation. We also
discuss the implication of this in the domain of computer aided visualisation.