Abstract
When we sculpt an object, what aspects of the object's geometrical information are depicted and to what extent? Furthermore, what can the statistical geometric properties of these depictions tell us about our mental representations of 3-D shape? Vision and haptics are largely responsible for our perception of 3-D shape. Past work has shown that our visual and haptic systems do not work congruently. Historically, art literature has suggested that what we perceive is often inharmonious with the real world. Previously, we examined the production of line drawings as a means to measure only visual perceptual ability. The present study examines visual and haptic perceptual differences using globally convex natural shaped 3-D objects. Sculpting objects relies on a direct mapping of three-dimensional information without the need for two-dimensional projection as in drawing. Participants were asked to sculpt objects based only on visual or haptic information. Two hands and normal vision were allowed for the actual sculpting. A stimulus set of 25 target objects was specified with a parametric range of statistical properties and manufactured with a 3-D printer. Each object's shape differed in spatial frequency and amplitude. The human-sculpted objects were scanned in 3-D and statistically compared to the original stimuli. An interaction was obtained between modality and ability with respect to object frequency. An object with a higher spatial frequency was harder to sculpt when limited to haptic input as compared to only visual input. The opposite was found for an object with a low spatial frequency. An associated visual-haptic comparison experiment, with the same stimulus set found complimentary results. The finding that perceptual abilities of our visual and haptic systems differ, suggests that each system creates a different non-universal mental representation of 3-D shape.
Treuhaft Fund for Art Technology, Skidmore College.