Abstract
Faces are more difficult to recognize than exemplars of other object categories, e.g., chairs, when viewed in reverse contrast (Subramaniam & Biederman, 1997). The visual mechanisms underlying this phenomenon, however, are not well understood. One possibility is that faces and objects rely on separate systems — one sensitive and the other insensitive to contrast reversal. Alternatively, the complex pigmentation patterns characteristic of faces may contribute to this phenomenon (Bruce & Langton, 1994; Liu et al., 1999; Kemp et al., 1996). To date there has not been a direct comparison of the same faces or objects with and without adbedo information (here, defined as the reflectance map of a surface independent of shading/shadows). We tested observers in a same/different sequential-matching task using grayscale images of the same face models with or without albedo information. Observers were instructed to base their judgments on face identity, ignoring any changes in contrast. On trials in which faces were shown in different contrasts (e.g., positive-negative), observers responded more slowly and less accurately as compared to trials in which both faces were shown in the same contrast. This decrement in performance was significantly greater for faces with albedo than for faces without albedo. These results indicate that reported differences between faces and objects across contrast reversal may be attributed to the presence or absence of informative surface information and not to stimulus category. This hypothesis was tested further by comparing the effect of contrast reversal on non-face objects with and without albedo. We conclude that contrast reversal disrupts the recognition of both faces and objects to a greater degree in the presence of informative albedo. That is, another putative dissociation between face and object processing can be accounted for by factors other than the object category per se.