"Modeling Expressions and Shapes of Human heads"
Period: January 2000 - July 2002
Type: European Research
The goal of MESH is no less than creating virtual heads that are indistinguishable from real ones in terms of their shape, texture, and dynamics (emotions and speech). Even for caricatures such properties ought to be convincing. We’ll learn from exemplars. Bringing to bear the latest in 3D acquisition, the project will capture detailed 3D dynamics of faces. Learned motions together with guidance by a ‘face space’ will allow us to animate models of existing and novel faces. Moreover, faces will not be handled in isolation but modeling will target complete heads, i.e. including ears, hair, eyes, teeth, tongue, neck and shoulders. MESH counts several companies among its contractors. They have direct plans for exploitation, both for production tools to be marketed and as tools to be used in productions of their own (for TV, internet and movies). Strengthening our partners’ position in these markets is a key objective
MIRALab’s involvement in the project is twofold. Firstly, MIRALab will provide set of face masks (generic faces) which suits specific requirement of the different resolutions i.e. for Internet to Movies through Broadcast quality. The face mask will be MPEG-4 compatible, so the animation specification will possible as a MPEG-4 stream.
MIRALab will develop complete set of hair styling and animation software for the project. Hair plays an important role in visual realism, which is the aim throughout the project.
Albert-Ludwigs University of Freiburg
Ecole Polytechnique Féderale de Lausanne, Virtual Reality Laboratory (VRLab)
Eidgenossische Technische Hoschule Zurich ETH
Imagination In Motion IIM
MIRALab, University of Geneva