Montoute T, Tiberghien G
Institut des Sciences Cognitives, UMR 5015 CNRS-VCBL, 67 boulevard Pinel, Bron, 69675, France.
Conscious Cogn. 2001 Dec;10(4):503-23. doi: 10.1006/ccog.2001.0524.
A common view in face recognition research holds that there is a stored representation specific to each known face. It is also posited that semantic or memory-based information cannot influence low-level face processing. The two experiments reported in this article investigate the nature of this representation and the flow of face information processing. Participants had to search for a particular primed face among other faces. In Experiment 1, the search was done in a context where distractors had either a different degree of fame or the same degree of fame. In Experiment 2, the target face was primed either with semantic information or without any information. Both experiments demonstrated that increasing the display set size lengthened face detection time. However, the lengthening was a function of face fame. The search context also had an effect on the slope of the famous face detection. The results are explained in terms of the idea that face representations are reconstructed and that high- and low-level information are integrated into the processing. The integration process is not a conscious one.
人脸识别研究中的一种常见观点认为，对于每个已知面孔都存在一种特定的存储表征。同时也假定语义或基于记忆的信息不会影响低级别的面部处理。本文报道的两项实验研究了这种表征的性质以及面部信息处理的流程。参与者必须在其他面孔中搜索特定的启动面孔。在实验1中，搜索是在干扰项具有不同知名度或相同知名度的背景下进行的。在实验2中，目标面孔要么用语义信息启动，要么不使用任何信息启动。两项实验均表明，增加显示集大小会延长面部检测时间。然而，延长是面孔知名度的函数。搜索背景也对著名面孔检测的斜率有影响。这些结果是根据面孔表征被重建以及高级和低级信息被整合到处理过程中的观点来解释的。整合过程不是一个有意识的过程。
