Abstract
Faces can be typically described both based on their perceptual attributes (e.g. face roundness, symmetry) and social attributes (e.g. how trustworthy is the face?). Which type of these judgements may improve our ability to recognize faces from new views we have not seen before? In a series of studies subjects learned unfamiliar frontal faces while making social, perceptual or no-judgments. We hypothesized that social judgments would generate a more abstract, view-invariant representation than perceptual judgements. During test subjects were presented with four face views from profile to frontal. Results show that social judgements improved recognition as compared to perceptual or no judgments, whereas perceptual judgements did not improve face recognition relative to no judgements. The benefit of socially-judged faces that were learned from frontal view was most prominent in three-quarter view of the same person but did not extend to profile view, which yielded low recognition rates in all conditions. In a follow-up study, faces were learned from frontal and ¾ views in study and were tested from the same and new views. Results show that presenting faces from multiple views did not improve recognition relative to single view presentation for both social and perceptual judgements. Finally, the benefit of social judgements over no judgements was not found for faces of other races, suggesting that the benefit of social judgments may be influenced by the social experience that we have with faces from our own race. Overall, our findings interestingly show that attending to social rather than perceptual aspects of faces may generate an abstract representation of the identity of the face, which particularly benefit face recognition from nearby views that we have not seen before.
Meeting abstract presented at VSS 2017