Representation of statistical properties
Representation of statistical properties
复制标题
DOI：
10.1016/s0042-6989(02)00596-5
复制
发表时间：
2003-02-01
期刊：
影响因子：
1.8
通讯作者：
Treisman, A
中科院分区：
文献类型：
--
作者：
Chong, SC;Treisman, A
日常场景通常包含多组相似的对象。知觉表征可以用统计描述符来概括这些。在确定了两个大小的心理平均数后，我们用12个不同大小的圆圈数组测量了判断平均数的阈值。它们的大小接近于均匀阵列和单个元素的大小，并且几乎不受曝光时间(50-1000ms)或记忆延迟(最多2s)的影响。它们在相同的分布范围内比在不同的分布范围(正态分布、均匀分布、双峰分布和均匀分布)下的精确度仅略高一些，这证实了受试者确实在计算平均尺寸。(C)2003爱思唯尔科学有限公司。保留所有权利。
Everyday scenes often contain sets of similar objects. Perceptual representations may summarize these with statistical descriptors. After determining the psychological mean of two sizes, we measured thresholds for judging the mean with arrays of 12 circles of heterogeneous sizes. They were close to those for the size of elements in homogeneous arrays and single elements, and were little affected by either exposure duration (50-1000 ms) or memory delays (up to 2s). They were only slightly more accurate within the same distribution than across different distributions (normal, uniform, two-peaks, and homogeneous), confirming that subjects were indeed averaging sizes. (C) 2003 Elsevier Science Ltd. All rights reserved.