WHAT REPRESENTS A FACE - A COMPUTATIONAL APPROACH FOR THE INTEGRATIONOF PHYSIOLOGICAL AND PSYCHOLOGICAL DATA

Empirical studies of face recognition suggest that faces might be stor ed in memory by means of a few canonical representations. The nature o f these canonical representations is, however, unclear. Although psych ological data show a three-quarter-view advantage, physiological studi es suggest profile and frontal views are stored in memory. A computati onal approach to reconcile these findings is proposed. The pattern of results obtained when different views, or combinations of views, are u sed as the internal representation of a two-stage identification netwo rk consisting of an autoassociative memory followed by a radial-basis- function network are compared. Results show that (i) a frontal and a p rofile view are sufficient to reach the optimal network performance; a nd (ii) all the different representations produce a three-quarter view advantage, similar to that generally described for human subjects. Th ese results indicate that although three-quarter views yield better re cognition than other views, they need not be stored in memory to show this advantage.