Gaze Driven Animation of Eyes

Abstract

We present a data driven model of eye movement, that includes movement of the globes, the periorbital soft tissues and eyelids and also the formation of wrinkles in the tissues. We describe a pipeline for measurement and estimation of tissue movement around the eyes using monocular high speed video capture. We use dense optical flow techniques to simultaneously estimate skin and globe motion, as well as high resolution texture images. Our methods are robust to transient occlusions. Finally, we present a system for interactive animation of eyes using a small number of animation parameters, including gaze. These parameters can be obtained from any source, such as keyframe animation or an actor's performance.