Haptic texture perception on 3D-printed surfaces transcribed from visual natural textures

Abstract

Humans have a sophisticated ability to discriminate surface textures by touch, which is valuable for discriminating materials. Conventional studies have investigated this ability by using stimuli with simple (lower-order) statistical structures. Nevertheless, the structure of natural textures can be much more complex, and the human brain can encode complex (higher-order) spatial structures at least when they are processed by the visual system. To see how much the tactile system can encode complex surface patterns, we 3D-printed textured surfaces based on visual images of natural scenes including leaves and stones and conducted a haptic texture discrimination experiment. The mean surface carving depths were equated among the patterns. The participants touched the patterns in three modes: passive scan, static touch, and vibration only. The results showed that the “photo” patterns, which were visually very different from one another, were nearly indiscriminable by touch regardless of the touching mode. This suggests that though human touch may be good at discriminating differences in simple spatial structures such as statistics about the amplitude spectrum, it is relatively insensitive to more complex spatial structures, possibly due to spatial and temporal summation of local signals. Although further investigation is necessary to fully understand spatial statistics relevant to tactile texture perception, directly comparing touch with vision by using the 3D printing technology is a promising research strategy.