Research map / Video / Video generation
OctWorld: Long-Range World-Consistent Video Generation with Octree-based 3D Mapping
Zelong Lv, Sicheng Xu, Jianfeng Xiang, Yue Dong, Ruicheng Wang, Yu Deng, Guangzhong Sun, Jiaolong Yang
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 |
|---|---|---|---|
| use OctMap memory instead of fixed-resolution TSDF (36 frames) (ablation) | OctWorld · Re10K reverse trajectory ablation | LPIPS: 0.1996 -> 0.1412 | Improved |
| use OctMap memory instead of fixed-resolution TSDF (36 frames) (ablation) | OctWorld · Re10K reverse trajectory ablation | Chamfer distance: 0.0345 -> 0.0252 | Improved |
| add OctMap spatial condition (ablation) | OctWorld · Re10K reverse trajectory ablation | LPIPS: 0.2128 -> 0.1412 | Improved |
| use TSDF instead of point cloud memory (16 frames) (ablation) | OctWorld · Re10K reverse trajectory ablation | Chamfer distance: 0.0916 -> 0.0556 | Improved |
| increase noise augmentation level from 0 to 3 (ablation) | OctWorld · WorldScore static | 3D consistency: 67.18 -> 71.67 | Improved |
| increase noise augmentation level from 0 to 3 (ablation) | OctWorld · Re10K reverse trajectory | LPIPS: 0.1412 -> 0.1633 | Hurt |
| use OctMap instead of fixed TSDF at voxel 0.002 (ablation) | OctWorld · Reprojection rendering | PSNR: 22.08 -> 23.47 | Improved |
| use TSDF with smaller voxel size (ablation) | OctWorld · Reprojection rendering | memory: 1.37 GB -> OOM at 0.001 | Hurt |
| use OctWorld instead of Gen3C | OctWorld · Re10K reverse trajectory | LPIPS: 0.1933 -> 0.1633 | Improved |
| use OctWorld instead of Voyager | OctWorld · WorldScore camera control | score: 46.60 -> 39.12 | Hurt |
| use OctWorld instead of Gen3C | OctWorld · WorldScore photometric consistency | score: 68.46 -> 77.01 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- AETHER — Aether: Geometric-Aware Unified World Modeling (ICCV 2025) · better on 3 of 3 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
- WorldStereo: Bridging Camera-Guided Video Generation and Scene Reconstruction via 3D Geometric Memories (CVPR 2026 · similarity 0.90)
- Endless World: Real-Time 3D-Aware Long Video Generation (CVPR 2026 · similarity 0.90)
- World-consistent Video Diffusion with Explicit 3D Modeling (CVPR 2025 · similarity 0.89)
- Video World Models with Long-term Spatial Memory (NeurIPS 2025 · similarity 0.88)
- WorldWeaver: Generating Long-Horizon Video Worlds via Rich Perception (NeurIPS 2025 · similarity 0.88)
- Diffusion4D: Fast Spatial-temporal Consistent 4D generation via Video Diffusion Models (NeurIPS 2024 · 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.
