Visualization of Facial Expression Deformation Applied to the Mechanism Improvement of Face Robot

Abstract

The static and dynamic realistic effects of the appearance are essential but challenging targets in the development of human face robots. Human facial anatomy is the primary theoretical foundation for designing the facial expressional mechanism in most existent human face robots. Based on the popular study of facial action units, actuators are arranged to connect to certain control points underneath the facial skin in prearranged directions to mimic the facial muscles involved in generating facial expressions. Most facial robots fail to generate realistic facial expressions because there are significant differences in the method of generating expressions between the contracting muscles and inner tissues of human facial skin and the wire pulling of a single artificial facial skin. This paper proposes a unique design approach, which uses reverse engineering techniques of three dimensional measurement and analysis, to visualize some critical facial motion data, including facial skin localized deformations, motion directions of facial features, and displacements of facial skin elements on a human face in different facial expressional states. The effectiveness and robustness of the proposed approach have been verified in real design cases on face robots.