Research map / 3D vision / 3D generation
Inferring Compositional 4D Scenes without Ever Seeing One
Ahmet Berke Gökmen, Ajad Chhatkuli, Luc Van Gool, Danda Pani Paudel
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 |
|---|---|---|---|
| add static/dynamic embeddings (ablation) | COM4D baseline · DeformingThings | IoU: 0.2018 -> 0.4034 | Improved |
| add static/dynamic embeddings (ablation) | COM4D baseline · DeformingThings | Chamfer Distance: 0.1525 -> 0.1284 | Improved |
| add static/dynamic embeddings (ablation) | COM4D baseline · 3D-FRONT | Chamfer Distance: 0.1668 -> 0.1044 | Improved |
| add Diffusion Forcing to baseline (ablation) | COM4D baseline · DeformingThings | IoU: 0.2018 -> 0.4271 | Improved |
| add Diffusion Forcing on top of embeddings (full model) (ablation) | COM4D · 3D-FRONT | Chamfer Distance: 0.1044 -> 0.0909 | Improved |
| add Diffusion Forcing on top of embeddings (full model) (ablation) | COM4D · DeformingThings | IoU: 0.4034 -> 0.4191 | No clear effect |
| add attention mixing at inference (ablation) | COM4D · CMU Panoptic point clouds | average Chamfer Distance (cm): 35.91 -> 7.42 | Improved |
| add attention mixing at inference (ablation) | COM4D · User study (spatial correctness, temporal coherence) | preference rate: 87% vs. 6.9% | Improved |
| replace frame-wise TripoSG with COM4D temporal modeling | TripoSG · DeformingThings4D | F-Score: 0.5179 -> 0.8388 | Improved |
| use COM4D instead of TripoSG | single-object 4D · Objaverse 4D subset | Chamfer Distance: 0.1107 -> 0.1205 | Hurt |
| use COM4D instead of MIDI on 3D scenes | MIDI · 3D-FRONT | Chamfer Distance: 0.1445 -> 0.0909 | Improved |
| use COM4D instead of PartCrafter on 3D scenes | PartCrafter · 3D-FRONT part IoU (lower better) | IoU: 0.0017 -> 0.0018 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- MIDI — MIDI: Multi-Instance Diffusion for Single Image to 3D Scene Generation (CVPR 2025) · better on 4 of 4 shared results
- PartCrafter — PartCrafter: Structured 3D Mesh Generation via Compositional Latent Diffusion Transformers (NeurIPS 2025) · better on 2 of 3 shared results
- L4GM — L4GM: Large 4D Gaussian Reconstruction Model (NeurIPS 2024) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Motion4D: Learning 3D-Consistent Motion and Semantics for 4D Scene Understanding (NeurIPS 2025 · similarity 0.88)
- Fused View-Time Attention and Feedforward Reconstruction for 4D Scene Generation (NeurIPS 2025 · similarity 0.88)
- X4D-SceneFormer: Enhanced Scene Understanding on 4D Point Cloud Videos through Cross-Modal Knowledge Transfer (AAAI 2024 · similarity 0.87)
- DreamScene4D: Dynamic Multi-Object Scene Generation from Monocular Videos (NeurIPS 2024 · similarity 0.87)
- CAT4D: Create Anything in 4D with Multi-View Video Diffusion Models (CVPR 2025 · similarity 0.87)
- IGGT4D: Streaming 4D Instance-Grounded Geometry Transformer (NeurIPS 2026 · similarity 0.87)
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.
