Division of Biology, California Institute of Technology, Nagoya University.
Psychol Sci. 2011 Feb;22(2):153-8. doi: 10.1177/0956797610393743. Epub 2010 Dec 21.
To analyze complex scenes efficiently, the human visual system performs perceptual groupings based on various features (e.g., color and motion) of the visual elements in a scene. Although previous studies demonstrated that such groupings can be based on a single feature (e.g., either color or motion information), here we show that the visual system also performs scene analyses based on a combination of two features. We presented subjects with a mixture of red and green dots moving in various directions. Although the pairings between color and motion information were variable across the dots (e.g., one red dot moved upward while another moved rightward), subjects' perceptions of the color-motion pairings were significantly biased when the randomly paired dots were flanked by additional dots with consistent color-motion pairings. These results indicate that the visual system resolves local ambiguities in color-motion pairings using unambiguous pairings in surrounds, demonstrating a new type of scene analysis based on the combination of two featural cues.
为了有效地分析复杂场景，人类视觉系统会根据场景中视觉元素的各种特征（例如颜色和运动）进行感知分组。尽管先前的研究表明，这种分组可以基于单个特征（例如颜色或运动信息），但我们在这里表明，视觉系统还可以基于两种特征的组合来进行场景分析。我们向受试者呈现了混合着红色和绿色点的画面，这些点以不同的方向移动。尽管点之间的颜色和运动信息的配对是变化的（例如，一个红点向上移动，而另一个红点向右移动），但是当随机配对的点被具有一致颜色-运动配对的额外点包围时，受试者对颜色-运动配对的感知会明显偏向于这些配对。这些结果表明，视觉系统使用周围环境中明确的配对来解决颜色-运动配对中的局部歧义，从而基于两种特征线索的组合来进行新的场景分析。
