World from Motion: Generative Dynamic Gaussian Reconstruction from Monocular Video
We present World from Motion, a method for generating freely renderable dynamic 3D Gaussian representations from monocular videos. Our approach conditions a video model on dense, pixel-aligned renderings that encode appearance, geometry, and 3D scene motion along both input and target camera trajectories to correct rendering artifacts and fill in missing regions from an initial reconstruction. To train this model, we construct a dataset of aligned multiview video pairs and dynamic 3DGS representations, with simulated artifacts characteristic of monocular reconstruction. At test time, we distil
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
- Linked via arxiv authorLiyuan Zhu →
World from Motion: Generative Dynamic Gaussian Reconstruction from Monocular Video
- Linked via arxiv authorShengyu Huang →
World from Motion: Generative Dynamic Gaussian Reconstruction from Monocular Video
- Linked via arxiv authorAmrita Mazumdar →
World from Motion: Generative Dynamic Gaussian Reconstruction from Monocular Video
- Linked via arxiv authorTianye Li →
World from Motion: Generative Dynamic Gaussian Reconstruction from Monocular Video
- Linked via arxiv authorZan Gojcic →
World from Motion: Generative Dynamic Gaussian Reconstruction from Monocular Video
- Linked via arxiv authorGordon Wetzstein →
World from Motion: Generative Dynamic Gaussian Reconstruction from Monocular Video
- Linked via arxiv authorIro Armeni →
World from Motion: Generative Dynamic Gaussian Reconstruction from Monocular Video
- Linked via arxiv authorShalini De Mello →
World from Motion: Generative Dynamic Gaussian Reconstruction from Monocular Video
- Linked via arxiv authorAlex Trevithick →
World from Motion: Generative Dynamic Gaussian Reconstruction from Monocular Video