|
Abstract:
|
Individuals with partial facial paralysis suffer severe psychological and social consequences as a result of the inability to create facial expressions such as smiles. Facial reanimation surgery offers hope for these individuals, but clinicians lack a detailed understanding of which features are most important for reanimating socially effective facial expressions. In this talk, I discuss how computer animated facial models and statistical learning methods can provide insight into how people perceive dynamic facial expressions of emotion. Our results reveal that a successful smile involves an intricate balance of mouth angle, smile extent, and teeth show combined with the correct amount of dynamic symmetry. These findings demonstrate the power of statistical learning methods for scientific discovery, and have practical implications for facial reanimation and rehabilitation.
|
