Presentation

<
Model See Model Do: Speech-Driven Facial Animation with Style Control
SessionAvatars
DescriptionModelSeeModelDo presents a speech-driven 3D facial animation method using a latent diffusion model conditioned on a reference clip to capture nuanced performance styles. A novel "style basis" mechanism extracts key poses to guide generation, achieving expressive, temporally coherent animations with accurate lip-sync and strong stylistic fidelity across diverse speech inputs.
Authors
Yifang Pan
University of Toronto
Jali Research, Canada
Karan Singh
University of Toronto
Luiz Gustavo Hafemann
Ubisoft
Event Type
Technical Paper
TimeWednesday, 13 August 202510:45am - 10:55am PDT
LocationWest Building, Rooms 220-222
ACM Digital Library Journal Papers' PDFs
Conference Papers' PDFs
Session Time & Location
Sunday, 10 August 20256:00pm - 8:45pm PDTWest Building, Ballroom AB
Wednesday, 13 August 202510:45am - 12:15pm PDTWest Building, Rooms 220-222
Recordings
Livestreamed
Recorded
Next PresentationNext PresentationTiny is not small enough: High quality, low-resource facial animation models through hybrid knowledge distillation
Similar Presentations
xADA: Controllable and Expressive Audio-Driven Animation
AutoKeyframe: Autoregressive Keyframe Generation for Human Motion Synthesis and Editing
Tiny is not small enough: High quality, low-resource facial animation models through hybrid knowledge distillation