Khang Byung-Geun, Zaidi Qasim
SUNY College of Optometry, 33 W 42nd Street, New York, NY 10036, USA.
Vision Res. 2002 Jan;42(2):211-26. doi: 10.1016/s0042-6989(01)00252-8.
The identification of objects, illuminants, and transparencies are probably the most important perceptual functions of color. This paper examines the effects of perceptual scission, image junctions, color adaptation, and color correlations on identification. Simulations of natural illuminants, materials, and filters were used in a forced-choice procedure to simultaneously measure thresholds for identifying filters and objects across illuminants, and discrimination thresholds within illuminants. In the vast majority of the cases, if observers could discriminate within illuminants they could identify across illuminants. Since results were similar for identical color distributions, whether transparency cues like X-junctions were present or not, the primary cues for color identification were systematic color shifts across illuminants. These color shifts can be well described by three-parameter affine transformations, and the parameters can be derived from differences and ratios of mean chromaticities. A strategy based on post-transformation color matching predicts generally accurate identification despite perceptible color shifts, and also provides plausible reasons for those few conditions where identification thresholds are significantly higher than discrimination thresholds.
物体、光源和透明物体的识别可能是颜色最重要的感知功能。本文研究了感知分裂、图像交界处、颜色适应和颜色相关性对识别的影响。在强制选择程序中使用了自然光源、材料和滤光片的模拟，以同时测量跨光源识别滤光片和物体的阈值，以及光源内的辨别阈值。在绝大多数情况下，如果观察者能够在光源内进行辨别，他们就能跨光源进行识别。由于对于相同的颜色分布，无论是否存在像X形交界处这样的透明线索，结果都是相似的，所以颜色识别的主要线索是跨光源的系统性颜色变化。这些颜色变化可以用三参数仿射变换很好地描述，并且这些参数可以从平均色度的差异和比率中推导出来。基于变换后颜色匹配的策略预测，尽管存在可感知的颜色变化，但识别通常是准确的，并且还为那些识别阈值明显高于辨别阈值的少数情况提供了合理的原因。
