Abstract
Little is known about the learning experiences that promote robust face recognition performance for highly familiar faces. This is because most face recognition research uses exclusively famous or exclusively unfamiliar faces as stimuli. In a previous study (Roark et al., in press), we found large gains in person recognition from low quality video, following multiple exposures to high resolution, frontal-view faces with neutral expressions. In the present experiment, we tested whether familiarizing participants with other types of face images (expressive faces and 3/4-view faces) would similarly improve performance. These kinds of images provide participants with faces that are potentially more socially engaging and offer better viewpoint information (cf. the 3/4 advantage) compared to neutral frontal-view faces. Participants learned faces from close-up images under controlled illumination and were tested with whole-body images under uncontrolled illumination. During learning we varied both the number of times participants viewed each face (1, 2, or 4 times) and the presentation type of the faces (neutral, smiling, or 3/4 view). We found equivalent and strong pure repetition effects for expressive, 3/4 view, and neutral frontal faces. These results suggest that the familiarity advantage we found previously is not dependent on a particular facial image type, but rather, is easily attainable from multiple exposures to any high-quality image. The repeated availability of a high-quality image seems to allow person recognition to tolerate large discrepancies in image format.
This work was supported by funding from TSWG awarded to AO'T and H.A.