Presentation

<
Reenact Anything: Semantic Video Motion Transfer Using Motion-Textual Inversion
SessionVery Vivid Videos
DescriptionReenact Anything introduces a unified framework for semantic motion transfer, covering applications from full-body and face reenactment to controlling the motion of inanimate objects and the camera. Thereby, motions are represented using text/image embeddings of an image-to-video diffusion model and are optimized based on a given motion reference video.
Authors
Manuel Kansy
ETH Zürich
Disney Research Studios
Jacek Naruniec
Disney Research Studios
Christopher Schroers
Disney Research Studios
Markus Gross
ETH Zürich
Disney Research Studios
Romann Weber
Disney Research Studios
Event Type
Technical Paper
TimeThursday, 14 August 202511:35am - 11:45am PDT
LocationWest Building, Rooms 118-120
ACM Digital Library Journal Papers' PDFs
Conference Papers' PDFs
Session Time & Location
Sunday, 10 August 20256:00pm - 8:45pm PDTWest Building, Ballroom AB
Thursday, 14 August 202510:45am - 12:15pm PDTWest Building, Rooms 118-120
Recordings
Not Livestreamed
Not Recorded
Next PresentationNext PresentationVery Vivid Videos - Interactive Discussion
Similar Presentations
Motion Inversion for Video Customization
Diffusion as Shader: 3D-aware Video Diffusion for Versatile Video Generation Control
VideoAnydoor: High-fidelity Video Object Insertion with Precise Motion Control