The paper proposes support-aware appearance baking via residual distillation from teacher anchors to produce consistent 3D Gaussian scenes from sparse 2D inputs while suppressing noise.
arXiv preprint arXiv:2312.05133 (2023)
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
years
2026 3roles
background 1polarities
background 1representative citing papers
Scaling transformer context with sparse attention and 3D-aware block routing improves feed-forward 3D reconstruction and inverse rendering, closing much of the quality gap with dense-view optimization.
F-RNG generates relightable 3D Gaussian splatting assets from sparse views via feed-forward distillation of IDM priors into an LRM without retraining the base models.
citing papers explorer
-
From Sparse and Imperfect 2D Anchors to Consistent 3D Gaussian Street Scenes: Support-Aware Appearance
The paper proposes support-aware appearance baking via residual distillation from teacher anchors to produce consistent 3D Gaussian scenes from sparse 2D inputs while suppressing noise.
-
LSRM: High-Fidelity Object-Centric Reconstruction via Scaled Context Windows
Scaling transformer context with sparse attention and 3D-aware block routing improves feed-forward 3D reconstruction and inverse rendering, closing much of the quality gap with dense-view optimization.
-
F-RNG: Feed-Forward Relightable Neural Gaussians
F-RNG generates relightable 3D Gaussian splatting assets from sparse views via feed-forward distillation of IDM priors into an LRM without retraining the base models.