Parametric Facial Animation for Affective Interaction Workflow for Avatar Retargeting

Virtual avatars/characters are essential in the interaction with virtual environments and 3D video games. While these avatars can be controlled by a bot or a human, they need a suitable model that allows emotion representation regardless of what or who is controlling them, since it increases the interaction and immersion. Provide facial expressions to any avatar require an animation process to emulate a particular gesture or emotion before to include it in a virtual environment. This process is specific to every single facial expression required in the avatar. This paper presents an initial approximation to contribute to the solve this problem with a workflow to apply a well-known parameterized face model called Candide. The proposal aims to adapt Candide into any avatar previously modeled through automatic morphological adjustment and texture mapping. This work includes the application of this workflow into six characters licensed under the Creative Commons copyright. We also present examples of facial emotion animation, as well as subjective assessment by five experts in audiovisual and animation production.