Silhouetted face profiles: A new methodology for face perception research

We present a new methodology for constructing face stimuli for psychological experiments based on silhouetted face profiles. Face silhouettes carry a number of theoretical and methodological advantages compared to more complex face stimuli and lend themselves to a simple yet powerful parameterization. In five behavioral studies, we show that face silhouettes are processed like regular face stimuli: They provide enough information for accurate gender judgments (Study 1), age estimations (Study 2), and reliable and cross-valid attractiveness ratings (Study 3). Furthermore, face silhouettes elicit an inversion effect (Study 4) and allow for remarkably accurate cross-identification with front-view photographs (Study 5). We then describe a shape-based parameterization that relies on a small set of landmark points and show that face silhouettes can be effectively represented in a 20-dimensional "silhouette face space" (Study 6). We show that in this physical space, distance from the center of the space corresponds to perceived distinctiveness (Study 7), confirming a key axiom in the formulation of the face space model. Finally, we discuss straightforward applications of the face silhouette methodology and address some limitations.