Research map / 3D vision / 3D generation
FantasyWorld: Geometry-Consistent World Modeling via Unified Video and 3D Prediction
Yixiang Dai, Fan Jiang, Chiyu Wang, Mu Xu, Yonggang Qi
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 geometry branch with bidirectional cross-attention (ablation) | FantasyWorld · WorldScore (small motion) | 3D consistency: 79.77 -> 83.31 | Improved |
| add geometry branch with bidirectional cross-attention (ablation) | FantasyWorld · WorldScore (large motion) | 3D consistency: 72.06 -> 74.83 | Improved |
| add geometry branch (ablation) | FantasyWorld · WorldScore (large motion) | photometric consistency: 56.98 -> 60.61 | Improved |
| add geometry branch (ablation) | FantasyWorld · WorldScore (small motion) | style consistency: 92.54 -> 94.22 | Improved |
| add geometry branch (ablation) | FantasyWorld · WorldScore (small motion) | object control: 37.33 -> 34.46 | Hurt |
| add geometry branch (ablation) | FantasyWorld · WorldScore (small motion) | content alignment: 43.31 -> 38.45 | Hurt |
| add geometry branch (ablation) | FantasyWorld · RealEstate10K 3DGS (VGGT init) | PSNR: 26.89 -> 28.24 | Improved |
| add geometry branch (ablation) | FantasyWorld · RealEstate10K 3DGS (VGGT init) | LPIPS: 0.17 -> 0.14 | Improved |
| initialize 3DGS from own predicted point cloud instead of VGGT (ablation) | FantasyWorld · RealEstate10K 3DGS | PSNR: 28.24 -> 26.54 | Hurt |
Closest papers by meaning
- DreamWorld: Geometry-Grounded Video Diffusion for 3D-Consistent World Modeling (ECCV 2026 · similarity 0.91)
- PanoWorld: Geometry-Consistent Panoramic Video World Modeling (NeurIPS 2026 · similarity 0.90)
- DriveDreamer-Policy: A Geometry-Grounded World-Action Model for Unified Generation and Planning (NeurIPS 2026 · similarity 0.89)
- 4DWorldBench: A Comprehensive Evaluation Framework for 3D/4D World Generation Models (CVPR 2026 · similarity 0.89)
- Video World Models with Long-term Spatial Memory (NeurIPS 2025 · similarity 0.88)
- 3D-VLA: A 3D Vision-Language-Action Generative World Model (ICML 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.
