Generalization to Novel Images in Upright and Inverted Faces

An image of a face depends not only on its shape, but also on the viewing position, illumination conditions, and facial expression. Any face recognition system must overcome the changes in face appearance induced by these factors. To assess the ability of human vision to generalize across changes in illumination and pose of faces, we studied the performance of subjects in a discrimination task with either upright or inverted faces. Subjects first learned to discriminate among images of three faces, taken under fixed viewing position and illumination. They were then tested on images of the same faces taken under all combinations of four illuminations and five viewing positions. For upright faces, we found remarkably good generalization to novel conditions. For inverted faces, the generalization to novel views was significantly worse, although the performance on the training images was similar in both cases. Our results indicate that at least some of the processes that support generaliza..