Wilson J Anthony, Switkes Eugene
Department of Psychology, University of California at Berkeley, California 94720, USA.
J Opt Soc Am A Opt Image Sci Vis. 2005 Oct;22(10):2169-81. doi: 10.1364/josaa.22.002169.
Using Glass patterns composed of isoluminant dots we have investigated the segregation and integration of chromatic information by the visual system. By measuring pattern detection when the chromaticities of the two elements forming a dot pair are varied (intradipole variation), we characterize integration at an early level of spatial processing. By measuring pattern detection for dot pairs where the within-pair chromaticity is the same but the among-pair chromaticities are varied (interdipole variation) we characterize integration and segregation for a more global, midlevel, spatial processing mechanism. Using isoluminant patterns in which all dots have the same chromaticity, we find that (i) detection thresholds are similar to those for luminance-defined dots, and (ii) an equivalent-contrast metric approximately equates thresholds for various chromaticities, including those along both the cardinal and the intermediate axes of an opponent-color space. When intradipole chromaticity is varied we observe that (i) the ability of visual mechanisms to extract oriented dot pairs decreases with increasing chromaticity differences, and (ii) average bandwidths are similar for cardinal and intermediate directions. For pattern detection with interdipole chromatic variation the visual system does not segregate noise dot pairs from correlated dot pairs on the basis of chromatic differences alone, and appears to integrate oriented dot pairs of differing chromaticities in forming a global percept, even for large color differences. Isoluminant Glass patterns with translational and concentric correlations give similar results. The results are compared with those obtained for contrast variation in luminance-defined Glass Patterns and are discussed in terms of current multistage models of color processing by the visual system.
我们使用由等亮度点组成的格拉斯图案，研究了视觉系统对颜色信息的分离和整合。通过测量形成点对的两个元素的色度发生变化时（偶极内变化）的图案检测，我们表征了空间处理早期阶段的整合情况。通过测量点对内色度相同但点对间色度不同的点对的图案检测（偶极间变化），我们表征了一种更全局、中级空间处理机制的整合和分离情况。使用所有点具有相同色度的等亮度图案，我们发现：（i）检测阈值与亮度定义的点的阈值相似；（ii）等效对比度度量大致使各种色度（包括沿对立颜色空间的主轴线和中间轴线的色度）的阈值相等。当偶极内色度发生变化时，我们观察到：（i）视觉机制提取定向点对的能力随色度差异的增加而降低；（ii）主方向和中间方向的平均带宽相似。对于偶极间色度变化的图案检测，视觉系统不会仅基于色度差异将噪声点对与相关点对分离，并且即使对于大的颜色差异，在形成全局感知时似乎也会整合不同色度的定向点对。具有平移和同心相关性的等亮度格拉斯图案给出了类似的结果。我们将这些结果与亮度定义的格拉斯图案中对比度变化所获得的结果进行了比较，并根据视觉系统当前的颜色处理多阶段模型进行了讨论。
