Honson Vanessa, Huynh-Thu Quan, Arnison Matthew, Monaghan David, Isherwood Zoey J, Kim Juno
Sensory Processes Research Laboratory, School of Optometry and Vision Science, University of New South Wales, Sydney, NSW, Australia.
Canon Information Systems Research Australia (CiSRA), Sydney, NSW, Australia.
Front Psychol. 2020 Mar 20;11:485. doi: 10.3389/fpsyg.2020.00485. eCollection 2020.
This study examined perceptual differentiation of specular from diffuse shading for the recovery of surface color and gloss. In Experiment 1, we parametrically varied the mesoscale relief height of globally planar surfaces, specular sharpness and the orientation of the surface relative to the light source. We obtained psychophysical matches for perceived color saturation and value (HSV), but also considered whether the main effects could be influenced by color space used when transforming data to perceptually-uniform CIE LCH space. Results revealed strong interactions between perceived color attributes and the lighting conditions, the structure of specular reflections, and surface relief. Declines in saturation were observed with increasing specular roughness (using an HSV color representation), but no similar decline was observed in chroma (using a CIE LCH color representation). Experiment 2 found strong negative correlations between perceived gloss and specular roughness. Perceived gloss also depended on mesoscopic relief height and orientation of the surface relative to the light source. Declines in perceived gloss moderately accounted for the variability in color saturation and value matches obtained in Experiment 1. We found information about perceived specular coverage could further improve the model's accountability of perceived color saturation and lightness (Experiment 3). These findings together suggest that perceived color saturation and color value depends on the visual system's ability to distinguish the underlying diffuse shading from specular highlights in images.
本研究考察了镜面反射与漫反射阴影的感知差异，以恢复表面颜色和光泽。在实验1中，我们参数化地改变了全局平面表面的中尺度起伏高度、镜面反射锐度以及表面相对于光源的方向。我们获得了感知颜色饱和度和明度（HSV）的心理物理匹配结果，同时还考虑了在将数据转换为感知均匀的CIE LCH空间时，所使用的颜色空间是否会影响主要效应。结果显示，感知颜色属性与光照条件、镜面反射结构和表面起伏之间存在强烈的相互作用。随着镜面反射粗糙度的增加（使用HSV颜色表示），饱和度下降，但在色度方面（使用CIE LCH颜色表示）未观察到类似的下降。实验2发现，感知光泽与镜面反射粗糙度之间存在强烈的负相关。感知光泽还取决于介观起伏高度和表面相对于光源的方向。感知光泽的下降适度地解释了实验1中获得的颜色饱和度和明度匹配的变异性。我们发现，关于感知镜面反射覆盖率的信息可以进一步提高模型对感知颜色饱和度和亮度的解释能力（实验3）。这些发现共同表明，感知颜色饱和度和颜色明度取决于视觉系统区分图像中潜在漫反射阴影与镜面反射高光的能力。
