Close

Presentation

ClipFace: Text-guided Editing of Textured 3D Morphable Models
SessionMaking Faces With Neural Avatars
DescriptionClipFace proposes novel self-supervised approach for text-guided editing of full-head textured 3D morphable model of faces by employing user-friendly language prompts to enable control of expressions as well as appearance of 3D faces. Our method enables new synthesis capabilities for out-of-domain textures and realistic expressions for 3DMMs, aligned with text.
Authors
Shivangi Aneja
Technical University of Munich
Justus Thies
Max Planck Institue for Intelligent Systems
Angela Dai
Technical University Munich
Matthias Niessner
Technical University Munich
Event Type
Technical Paper
TimeThursday, 10 August 202311:36am - 11:46am PDT
LocationPetree Hall D
Interest Areas
Research & Education
Recordings
Livestreamed
Recorded
Keywords
Animation/Simulation
Modeling
Video
Registration Categories
FC
FCS
V
VS
EFC
Next PresentationNext Presentation
Making Faces With Neural Avatars - Q&A and Interactive Discussion