VISUAL SPEECH SYNTHESIS FROM 3D VIDEO

In this paper we introduce a process for the synthesis of visual speech from captured 3D video. Animation from real speech is performed by path optimisation over a graph structure containing dynamic phonetic units.