Research map / Video / Video generation
One4D: Unified 4D Generation and Reconstruction via Decoupled LoRA Control
Zhenxing Mi, YUXIN WANG, Dan Xu
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 |
|---|---|---|---|
| jointly train generation and reconstruction with DLC+UMC vs recon-only MonST3R/CUT3R | One4D · Sintel depth | delta<1.25: 58.5 / 62.0 -> 70.4 | Improved |
| unified G&R model vs recon-only Geo4D-ref | One4D · Sintel depth | Abs Rel: 0.205 -> 0.273 | Hurt |
| unified model on Bonn vs MonST3R | One4D · Bonn depth | Abs Rel: 0.063 -> 0.092 | Hurt |
| camera estimation from One4D pointmaps | One4D · Sintel/TUM camera | ATE: Sintel 0.108 (MonST3R) vs 0.213; TUM 0.063 vs 0.129 | Mixed |
| replace spatial-concat 4DNeX with One4D (DLC) | One4D · single-image-to-4D VBench | Dynamic degree: 25.6% -> 55.7% | Improved |
| replace 4DNeX with One4D | One4D · single-image-to-4D VBench | I2V consistency: 98.7% -> 97.8% | No clear effect |
| user study One4D vs 4DNeX | One4D · 4D generation | 4D preference: 10.0% -> 90.0% | Improved |
| reduce observed frames to 50% (ablation) | One4D · Sintel sparse-frame | Abs Rel: 0.273 -> 0.314 | No clear effect |
| reduce observed frames to 5% (ablation) | One4D · Sintel sparse-frame | Abs Rel: 0.273 -> 0.641 | Hurt |
| use CFG=4 instead of 6 (ablation) | One4D · Sintel depth | Abs Rel: 0.273 -> 0.257 | No clear effect |
| train longer (1K to 5.5K steps) (ablation) | One4D · Sintel depth | Abs Rel: 0.331 -> 0.273 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- DepthCrafter — DepthCrafter: Generating Consistent Long Depth Sequences for Open-world Videos (CVPR 2025) · better on 1 of 2 shared results
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
- EG4D: Explicit Generation of 4D Object without Score Distillation (ICLR 2025 · similarity 0.88)
- Restage4D: Reanimating Deformable 3D Reconstruction from a Single Video (NeurIPS 2025 · similarity 0.88)
- Instant4D: 4D Gaussian Splatting in Minutes (NeurIPS 2025 · similarity 0.88)
- CAT4D: Create Anything in 4D with Multi-View Video Diffusion Models (CVPR 2025 · similarity 0.87)
- 4D-LRM: Large Space-Time Reconstruction Model From and To Any View at Any Time (NeurIPS 2025 · similarity 0.87)
- Consistent4D: Consistent 360° Dynamic Object Generation from Monocular Video (ICLR 2024 · 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.
