Research map / Video / Video-language understanding
OmniWorld: A Multi-Domain and Multi-Modal Dataset for 4D World Modeling
Yang Zhou, Yifan Wang, Jianjun Zhou, Wenzheng Chang, Haoyu Guo, Zizun Li, Kaijing Ma, Xinyue Li, Yating Wang, Haoyi Zhu, Mingyu Liu, Dingning Liu and 7 more
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 |
|---|---|---|---|
| fine-tune on OmniWorld | DUSt3R · Sintel monocular depth | Abs Rel: 0.488 -> 0.370 | Improved |
| fine-tune on OmniWorld | DUSt3R · Bonn monocular depth | Abs Rel: 0.139 -> 0.067 | Improved |
| fine-tune on OmniWorld | DUSt3R · NYU-v2 monocular depth | Abs Rel: 0.081 -> 0.089 | Hurt |
| fine-tune on OmniWorld | CUT3R · Bonn monocular depth | Abs Rel: 0.058 -> 0.075 | Hurt |
| fine-tune on OmniWorld | CUT3R · KITTI monocular depth | Abs Rel: 0.097 -> 0.087 | Improved |
| fine-tune on OmniWorld | CUT3R · Sintel video depth (scale&shift) | Abs Rel: 0.537 -> 0.314 | Improved |
| fine-tune on OmniWorld | DUSt3R · Sintel video depth (scale) | Abs Rel: 0.652 -> 0.512 | Improved |
| fine-tune on OmniWorld | DUSt3R · KITTI video depth (scale) | delta<1.25: 0.814 -> 0.800 | Hurt |
| fine-tune on OmniWorld | Reloc3r · DynPose-100K relative pose | AUC@5: 6.9 -> 14.4 | Improved |
| fine-tune on OmniWorld | CUT3R · Sintel camera pose | RPE rot: 0.627 -> 0.651 | Hurt |
| fine-tune on OmniWorld | AC3D · OmniWorld-Game benchmark | TransErr: 6.2788 -> 4.1428 | Improved |
| fine-tune on OmniWorld | AC3D · RealEstate10K | TransErr: 3.4433 -> 2.8648 | Improved |
Closest papers by meaning
- 4DWorldBench: A Comprehensive Evaluation Framework for 3D/4D World Generation Models (CVPR 2026 · similarity 0.89)
- DynamicVerse: A Physically-Aware Multimodal Framework for 4D World Modeling (NeurIPS 2025 · similarity 0.89)
- GameVerse: A Minute-Scale Gameplay Dataset for Long-Horizon Interactive World Modeling (NeurIPS 2026 · similarity 0.88)
- Learning 4D Embodied World Models (ICCV 2025 · similarity 0.88)
- stable-worldmodel: A Platform for Reproducible World Modeling Research and Evaluation (NeurIPS 2026 · similarity 0.86)
- LoopNav: Benchmarking Spatial Consistency in World Models (NeurIPS 2026 · 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.
