Abstract
The image of an object depends on many factors: the object's shape, surface material and the light field. Thus, it would be sensible for the visual system to jointly estimate all three, leading to the hypothesis that varying cues to shape should impact perceived gloss and vice versa. We present several sets of experiments in which participants judged gloss and/or shape. In one experiment, objects were rendered in one light field (taken from the Southampton-York Natural Scenes dataset of high dynamic range, natural light fields), and sometimes placed with a background from a different or altered light field. We found modest but systematic effects of the incongruent light field suggesting that observers are sensitive to the context in which objects are presented when making judgments of surface gloss. When participants rated gloss and shape against physical scales (real objects varying in gloss or bumpiness), perceived gloss increased with rendered gloss and bumpiness. Although perceived bumpiness was largely independent of physical gloss, bumpiness was increasingly underestimated as depth increased. This supports the hypothesis that shape and gloss are jointly estimated: underestimation of shape is coupled with overestimation of gloss, consistent with the effects of these two variables on images of physical objects. Finally, we modified perceived shape by allowing participants to touch an object (with a haptic force-feedback device). Stretching the object along the line of sight did not affect perceived gloss, but changing the qualitative shape of a bi-stable stimulus had substantial effects: matte objects became glossy or vice versa.