Presentation

<
Sparse-to-Complete: From Sparse Image Captures to Complete 3D Scenes
SessionDynamic Scenes
DescriptionWe introduce S2C-3D, a novel sparse-view 3D reconstruction framework for high-fidelity, complete scene reconstruction. It features three components: a specialized diffusion model for scene-specific image restoration, a training-free view-consistency conditioned sampling process in the diffusion model for refined Gaussian optimization, and a camera trajectory planning scheme for comprehensive scene coverage.
Authors
Yiyang Shen
State Key Lab of CAD and CG, Zhejiang University
Yin Yang
The University of Utah
Kun Zhou
State Key Lab of CAD and CG, Zhejiang University
Tianjia Shao
State Key Lab of CAD and CG, Zhejiang University
Event Type
Technical Paper
TimeTuesday, 21 July 202611:25am - 11:35am PDT
LocationRoom 408 B
Digital Library PDF
Session Time & Location
Sunday, 19 July 20266:00pm - 8:45pm PDTHall K
Tuesday, 21 July 202610:45am - 12:15pm PDTRoom 408 B
Keywords
Geometry
Modeling
Rendering
Registration Categories
Full Conference Supporter
Full Conference
Next PresentationNext PresentationArticulate That Object Part (ATOP): 3D Part Articulation from Text and via Motion Personalization
Similar Presentations
VidSplat: Gaussian Splatting Reconstruction with Geometry-Guided Video Diffusion Priors
Mix3R: Mixing Feed-forward Reconstruction and Generative 3D Priors for Joint Multi-view Aligned 3D Reconstruction and Pose Estimation
GeoQuery: Geometry-Query Diffusion for Sparse-View Reconstruction