It has recently been shown that an increase of the relief height of a glossy surface positively correlates with the perceived level of gloss (Y.-H. Ho, M. S. Landy, & L. T. Maloney, 2008). In the study presented here we investigated whether this relation could be explained by the finding that glossiness perception correlates with the skewness of the luminance histogram (I. Motoyoshi, S. Nishida, L. Sharan, & E. H. Adelson, 2007). First, we formally derived a general relation between the depth range of a Lambertian surface, the illumination direction and the associated image intensity transformation. From this intensity transformation we could numerically simulate the relation between relief stretch and the skewness statistic. This relation predicts that skewness increases with increasing surface depth. Furthermore, it predicts that the correlation between skewness and illumination can be either positive or negative, depending on the depth range. We experimentally tested whether changes in the depth range and illumination direction alter the appearance. We indeed find a convincingly strong illusory gloss effect on stretched Lambertian surfaces. However, the results could not be fully explained by the skewness hypothesis. We reinterpreted our results in the context of the bas-relief ambiguity (P. N. Belhumeur, D. J. Kriegman, & L. Yuille, 1999) and show that this model qualitatively predicts illusory highlights on locations that differ from actual specular highlight locations with increasing illumination direction.

Abstract-Prior studies have shown that local surface orientation is a dominant source of information for haptic curvature perception in static conditions. We show that this dominance holds for dynamic touch, just as was shown earlier for static touch. Using an apparatus specifically developed for this purpose, we tested this hypothesis by providing observers with two independently controlled sources of geometric information. The robotic-like apparatus could accurately control the position of a contact surface independently from its orientation in space, while allowing subjects to freely and actively explore virtual shapes in the lateral direction. In the first experiment, we measured discrimination thresholds for the two types of shape information and compared the discrimination of real shapes to that of virtual shapes. The results confirmed the dominance of local surface orientation. We propose a model that predicts cue dominance for different scales of exploration. In the second experiment we investigated whether a virtual curved surface felt as curved as a real curved surface. We found that observers did not systematically judge either of the two kinds of stimuli to be more curved than the other. More importantly, we found that points of subjective curvedness were not influenced by the availability of height information.

Abstract. We demonstrate the influence of picture size on haptic recognition and exploratory behaviour. The stimuli were raised-line drawings of everyday objects. Participants were instructed to think aloud during haptic exploration of the pictures. We measured the delay between initial correct speculation and final correct response. The results indicate that picture size influences accuracy but not response latency: large drawings are recognised more often but not faster. By analysing video recordings of the experiment we found that two-handed exploration increases when picture size increases and that, on average, 83% of the exploration time involves the use of two hands. The thinking-aloud data showed that the average time difference between the initial correct speculation and final correct response amounted to 23% of the total reaction time. We discuss our results with respect to the design of tactile aids and the ecological validity of single-finger exploration.

To understand the key image features that we use to infer the glossiness of materials, we analyzed the pictorial shortcuts used by 17th century painters to imitate the optical phenomenon of specular reflections when depicting grapes. Gloss perception of painted grapes was determined via a rating experiment. We computed the contrast, blurriness, and coverage of the grapes' highlights in the paintings' images, inspired by Marlow and Anderson (2013). The highlights were manually segmented from the images, and next the features contrast, coverage, and blurriness were semiautomatically quantified using self-defined algorithms. Multiple linear regressions of contrast and blurriness resulted in a predictive model that could explain 69% of the variance in gloss perception. No effect was found for coverage. These findings are in agreement with the instructions to render glossiness of grapes contained in a 17th century painting manual (Beurs, 1692/in press), suggesting that painting practice embeds knowledge about key image features that trigger specific material percepts.

The difficulty that observers experience when trying to identify a raised line drawing by touch is still largely unexplained. In this article, we show that observers who are unable to haptically identify a raised line drawing are suddenly able to do so after they have sketched on paper what they have in their mind. We conducted three experiments: first of all we show that this effect is robust; in the second experiment, we show that identification-after-sketching is caused by visual inspection of the sketch, and not caused by feedback in general; and in the third we show that sketches which were identified by the observers who produced them, were also identified by completely naive viewers. These experiments demonstrate that during raised line drawing identification the mental capacities required to interpret the stimulus seem to be inadequate: although enough pictorial information was present to produce a sketch which could even be identified by naive viewers, the stimulus could not be identified by haptic and mental processing alone. Furthermore, we investigated whether increasing the haptic perceptual field by using two hands instead of one hand had an influence on identification performance. We did indeed find that using two hands significantly increased identification. We use both results to discuss the underlying mechanisms of haptic raised line drawing identification.