Making faces: creating three-dimensional parameterized models of facial expression

Previous studies of the perceptual processing and dynamics of emotional expression in faces have been limited by the lack of realistic yet controlled stimuli. The present work offers researchers a method for creating such stimuli for exploring these phenomena. We describe the creation of the stimuli and a series of experiments testing the validity of these stimuli with respect to emotional expressions in humans. Participants evaluated synthesized facial images and standardized photographs of six basic emotional expressions for intensity and accuracy of perceived emotion. Comparisons of these measures were qualitatively similar for synthesized and photographed faces. A manipulation of the magnitude of the synthesized expressions yielded a significant effect on the perceived intensity of expression. In a subsequent multidimensional scaling study, no systematic differences were uncovered in the derived configurations of the synthesized expressions and the photographs. These results are discussed in the context of possible future research applications.