An Optimized Marker Layout for 3D Facial Motion Capture

Abstract

Facial motion capture (Facial Mocap) has gained increasing attention from researchers and professionals from different areas of interest, including entertainment, face-to-face communication, and training. Facial Mocap allows straightforward capture of dynamic behavior of the face from live action, providing data that can be used to drive realistic animation of a 3D virtual face. Facial Mocap is an advantageous alternative to the direct and laborious manipulation of the face model. In particular, marker-based mocap technique acquires three-dimensional facial points trajectories by tracking markers fixed on the face of an actor. However, despite the existence of several empirical facial marker layouts, the ideal positioning of the markers is still an open question. This paper presents an optimization technique to calculate the quantity and positioning facial markers and establish their influences on the polygon mesh based on the correlation of markers in a dense layout. The technique generates an optimized marker layout discarding unnecessary markers and positioning the remaining ones.