Research map / 3D vision / 3D generation
Unsupervised object-centric video generation and decomposition in 3D
Paul Henderson, Christoph H. Lampert
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use voxel objects instead of mesh objects (ablation) | O3V · traffic | MRE: 1.260 -> 0.201 | Improved |
| Use voxel objects instead of mesh objects (ablation) | O3V · traffic | FVD: 802.6 -> 386.6 | Improved |
| Use mesh objects instead of voxel objects (ablation) | O3V · rooms | FID: 121.0 -> 118.4 | Improved |
| Use voxel objects instead of mesh objects (ablation) | O3V · rooms | 3D detection AP: 0.250 -> 0.332 | Improved |
| Use voxel objects instead of mesh objects (ablation) | O3V · traffic | 3D detection AP: 0.060 -> 0.035 | Hurt |
| Use an explicit 3D scene-level prior rather than independent 2D object sprites | Video generation · rooms | FVD: 1421.6 -> 341.8 | Improved |
| Use an explicit 3D scene-level prior rather than independent 2D object sprites | Video generation · traffic | FVD: 620.4 -> 386.6 | Improved |
| Use the proposed video model instead of GENESIS | Video/frame generation · rooms | FID: 96.9 -> 121.0 | Hurt |
| Use the proposed model instead of SCALOR | Instance segmentation · rooms | SC, per-frame: 0.779 -> 0.652 | Hurt |
| Use the proposed model instead of SCALOR | Instance segmentation · rooms | SC, tracking: 0.509 -> 0.655 | Improved |
Closest papers by meaning
- DreamScene4D: Dynamic Multi-Object Scene Generation from Monocular Videos (NeurIPS 2024 · similarity 0.90)
- Robust Multi-Object 4D Generation for In-the-wild Videos (CVPR 2025 · similarity 0.88)
- Believing is Seeing: Unobserved Object Detection using Generative Models (CVPR 2025 · similarity 0.87)
- Generative Omnimatte: Learning to Decompose Video into Layers (CVPR 2025 · similarity 0.87)
- GEN3C: 3D-Informed World-Consistent Video Generation with Precise Camera Control (CVPR 2025 · similarity 0.86)
- Generative Video Propagation (CVPR 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
