Research map / Video / Video generation
WorldReel: 4D Video Generation with Consistent Geometry and Motion Modeling
Shaoheng Fang, Hanwen Jiang, Yunpeng Bai, Niloy J. Mitra, Qixing Huang
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 |
|---|---|---|---|
| WorldReel vs GeoVideo trained on same data | CogVideoX-5B-I2V · SpatialVid general motion | FVD: 371.3 -> 336.1 | Improved |
| WorldReel vs GeoVideo on complex motion | CogVideoX-5B-I2V · SpatialVid complex motion | FVD: 409.9 -> 394.2 | Improved |
| joint geometry and motion modeling raises dynamic degree | WorldReel · SpatialVid complex motion | dynamic degree: 0.79 -> 1.00 | Improved |
| WorldReel depth vs GeoVideo | WorldReel · generated 4D geometry | depth log-RMSE: 0.353 -> 0.287 | Improved |
| add geo-motion augmented latent (ablation) | WorldReel · SpatialVid complex motion | FVD: 452.8 -> 394.2 | Improved |
| add geo-motion augmented latent (ablation) | WorldReel · generated 4D geometry | depth log-RMSE: 0.352 -> 0.287 | Improved |
| add joint training with regularization (ablation) | WorldReel · SpatialVid complex motion | FVD: 411.8 -> 394.2 | Improved |
| add joint training with regularization (ablation) | WorldReel · generated 4D geometry | depth log-RMSE: 0.399 -> 0.287 | Improved |
| joint training with regularization on RGB-only model vs plain finetuning (ablation) | CogVideoX-5B-I2V · SpatialVid complex motion | FVD: 437.0 -> 452.8 | Hurt |
| freeze temporal DPT heads during joint training (ablation) | WorldReel · SpatialVid complex motion | FVD: 394.2 -> 382.3 | Mixed |
| freeze temporal DPT heads during joint training (ablation) | WorldReel · SpatialVid general motion | FID: 36.58 -> 38.02 | Hurt |
| WorldReel vs 4DNeX on motion | WorldReel · SpatialVid general motion | dynamic degree: 0.03 -> 0.73 | Improved |
Closest papers by meaning
- 4Real: Towards Photorealistic 4D Scene Generation via Video Diffusion Models (NeurIPS 2024 · similarity 0.89)
- World from Motion: Generative Dynamic Gaussian Reconstruction from Monocular Video (NeurIPS 2026 · similarity 0.89)
- PanoWorld: Geometry-Consistent Panoramic Video World Modeling (NeurIPS 2026 · similarity 0.89)
- GenXD: Generating Any 3D and 4D Scenes (ICLR 2025 · similarity 0.89)
- Diffusion4D: Fast Spatial-temporal Consistent 4D generation via Video Diffusion Models (NeurIPS 2024 · similarity 0.88)
- WorldWeaver: Generating Long-Horizon Video Worlds via Rich Perception (NeurIPS 2025 · similarity 0.88)
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.
