摘要: Automatic facial action unit (AU) detection is a research topic that finds many applications in behavioral science and human computer interaction. The AU performance 2D images are maturing but not yet adequate. In this study, we develop method to detect AUs 3D show its superiority vis-a-vis 2D. data modality curvature map, which obtained by conformal mapping of surface data. Since comparisons run on with the same algorithms, any bias could be induced precluded. We address also choice generative versus discriminative classifiers, consider 2D-3D fusion.