Abstract
Material surfaces have enormous visual information regarding their qualities. On the other hand, human observers can recognize material categories at a glance. This suggests that the human visual system can immediately extract small but enough amount of information to perceive material categories. To investigate the pieces of information that support the material perception, we conducted two experiments about surface quality evaluation and material discrimination.
The stimuli were presented on a gamma-corrected CRT display. They were photographs of the real material samples we prepared. The samples were size- and shape-controlled fabrics, glasses, leathers, metals, plastics, stones, and woods. More than ten samples (11 to 13) were prepared for each material. In Experiment 1, participants observed the stimulus and evaluated nine material features, which were both visual and non-visual, such as glossiness, transparency, roughness and so on. In Experiment 2, two photographs were presented simultaneously, and the task was to answer whether they belonged to the same material category or not, as quickly as possible.
Multi-dimensional scaling applied to the results of Experiment 1 suggested that the nine material features can be represented by about three components. Multi regression analysis applied to the results of Experiment 2 suggested that the reaction time can be explained mostly by three components almost the same as those of Experiment 1. At the same time, both analyses showed that all the nine features contributed to the material categorization. It was also suggested that not only visual features but also non-visual features contribute to the discrimination of material category. On the other hand, our analyses also revealed that other material features in addition to the nine features should be introduced to explain the reaction time for material discrimination.
Meeting abstract presented at VSS 2012