2026 年 38 巻 3 号 p. 143-159
Human perception of three-dimensional (3D) shape from two-dimensional images is strongly influenced by illumination. Previous studies have shown that changes in lighting direction can cause observers to misjudge identical 3D shapes as different under monocular and static viewing conditions. Such misperception may lead to practical problems when 3D computer graphics are used to present design proposals or final products in professional practice. This study investigates whether surface properties can reduce illumination-induced misperception of 3D shape. Focusing on specular reflection and shading strength, we conducted the shape re-identification task using a two-alternative forced-choice paradigm under multiple surface conditions. The results showed improved performance with specular reflection and weaker shading. In addition, to examine whether observers relied on subjective image similarity rather than perceived shape and the involvement of higher-level visual mechanisms, we performed a supplementary experiment and algorithmic simulations based on objective image quality metrics.