Wolfe J M
Harvard Medical School, Cambridge, MA 02138.
Vision Res. 1994 May;34(9):1187-95. doi: 10.1016/0042-6989(94)90300-x.
In standard visual search experiments, subjects look for a target item among distractor items. Targets and distractors are generally presented as isolated entities on a blank background. Do theories based on results of such experiments have any application to the real world of spatially continuous stimuli where "items" are not so conveniently delineated? Experiment 1 examines search for conjunctions of color with form and with orientation in novel stimuli designed to resemble aerial views. As with artificially isolated items, conjunction search with these complex stimuli is highly efficient, holding out the promise that models of visual search based on laboratory data can be extended to real-world situations. Experiment 2 shows that some "serial" searches are made significantly less efficient when items are embedded in these specific naturalistic backgrounds. This suggests that even classic serial searches rely on some parallel processing--the processing required to locate the "items" to be serially searched.
在标准的视觉搜索实验中，受试者要在干扰项中寻找目标项。目标项和干扰项通常作为孤立的实体呈现在空白背景上。基于此类实验结果的理论能否应用于空间连续刺激的现实世界？在这个世界中，“项目”并非如此方便地被界定。实验1研究了在设计成类似鸟瞰图的新型刺激中对颜色与形状及方向的组合的搜索。与人为孤立的项目一样，对这些复杂刺激进行组合搜索效率很高，这表明基于实验室数据的视觉搜索模型有望扩展到现实世界的情况。实验2表明，当项目嵌入这些特定的自然主义背景中时，一些“串行”搜索的效率会显著降低。这表明，即使是经典的串行搜索也依赖于一些并行处理——即定位要进行串行搜索的“项目”所需的处理。
