A single-pass 3D-native network upgrades low-quality Gaussian Splatting assets with local geometric anchors, delivering SOTA fidelity on a new benchmark at up to 10^5× the speed of optimization pipelines.
arXiv preprint arXiv:2411.06390 (2024)
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
fields
cs.CV 3years
2026 3representative citing papers
L2D2-GS reformulates generalizable dynamic Gaussian reconstruction as iterative optimization with a self-supervised densification policy and geometric regularization, claiming SOTA fidelity and zero-shot generalization on PandaSet and Waymo with fewer primitives.
GaussTrace constructs directed provenance graphs for 3DGS models via attribute-wise statistical profiling, hypothesis-driven editing simulations, and LLM-based evidence reasoning without training or edit histories.
citing papers explorer
-
AnchorSplat: Fast and Structure Consistent Detail Synthesis for Gaussian Splatting
A single-pass 3D-native network upgrades low-quality Gaussian Splatting assets with local geometric anchors, delivering SOTA fidelity on a new benchmark at up to 10^5× the speed of optimization pipelines.
-
L2D2-GS: Learning to Densify for Feedforward Dynamic Gaussian Scene Reconstruction
L2D2-GS reformulates generalizable dynamic Gaussian reconstruction as iterative optimization with a self-supervised densification policy and geometric regularization, claiming SOTA fidelity and zero-shot generalization on PandaSet and Waymo with fewer primitives.
-
GaussTrace: Provenance Analysis of 3D Gaussian Splatting Models with Evidence-based LLM Reasoning
GaussTrace constructs directed provenance graphs for 3DGS models via attribute-wise statistical profiling, hypothesis-driven editing simulations, and LLM-based evidence reasoning without training or edit histories.