Collishaw S M, Hole G J
School of Cognitive and Computing Sciences, University of Sussex, Falmer, Brighton BN1 9QH, UK.
Perception. 2000;29(8):893-909. doi: 10.1068/p2949.
Previous research suggests that face recognition may involve both configurational and piecemeal (featural) processing. To explore the relationship between these processing modes, we examined the patterns of recognition impairment produced by blurring, inversion, and scrambling, both singly and in various combinations. Two tasks were used: recognition of unfamiliar faces (seen once before) and recognition of highly familiar faces (celebrities). The results provide further support for a configurational-featural distinction. Recognition performance remained well above chance if faces were blurred, scrambled, inverted, or simultaneously inverted and scrambled: each of these manipulations disrupts either configurational or piecemeal processing, leaving the other mode available as a route to recognition. However, blurred/scrambled and blurred/inverted faces were recognised at or near chance levels, presumably because both configurational processing and featural processing were disrupted. Similar patterns of effects were found for both familiar and unfamiliar faces, suggesting that the relationship between configurational and featural processing is qualitatively similar in both cases.
先前的研究表明，人脸识别可能涉及构型加工和零碎（特征）加工。为了探究这些加工模式之间的关系，我们研究了模糊、倒置和打乱处理单独以及各种组合所产生的识别损伤模式。我们使用了两项任务：识别不熟悉的面孔（之前只见过一次）和识别非常熟悉的面孔（名人）。研究结果为构型-特征区分提供了进一步的支持。如果面孔被模糊、打乱、倒置或同时倒置并打乱，识别表现仍远高于随机水平：这些操作中的每一种都会破坏构型加工或零碎加工，使另一种加工模式成为识别的途径。然而，模糊/打乱和模糊/倒置的面孔在随机水平或接近随机水平被识别，可能是因为构型加工和特征加工都受到了干扰。在熟悉面孔和不熟悉面孔中都发现了类似的效应模式，这表明在这两种情况下，构型加工和特征加工之间的关系在性质上是相似的。
