Research map / 3D vision / 3D generation
Doe-2: 3D Representation World Model for Unified Driving Scene Forecasting
Dong Zhuo, Wenzhao Zheng, Sicheng Zuo, Yuanhui Huang, Siming Yan, Lu Hou, Jie Zhou, Jiwen Lu
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 3-layer DiT head to pure AR model (ablation) | Doe-2 · nuScenes generation | gFID: 29.93 -> 10.48 | Improved |
| replace pure DiT with AR+DiT head (ablation) | Doe-2 · nuScenes generation | FPS: ~1.9 -> ~4.0 | Improved |
| replace pure DiT with AR+DiT head (ablation) | Doe-2 · nuScenes generation | gFID: 16.25 -> 10.48 | Improved |
| compress latent dimension from 768 to 64 (ablation) | Doe-2 · SurroundOcc generation | mIoU: 8.24 -> 12.04 | Improved |
| compress latent dimension from 768 to 64 (ablation) | Doe-2 · SurroundOcc | FPS: ~2.3 -> ~4.0 | Improved |
| compress latent dimension from 64 to 32 (ablation) | Doe-2 · SurroundOcc generation | mIoU: 12.04 -> 9.33 | Hurt |
| raise latent dimension from 64 to 768 (ablation) | Doe-2 · SurroundOcc reconstruction | mIoU: 19.65 -> 20.02 | Improved |
| lower KL weight from 0.01 to 0.00001 (ablation) | Doe-2 · SurroundOcc generation | mIoU: 8.92 -> 12.04 | Improved |
| lower KL weight from 0.00001 to 0.000001 (ablation) | Doe-2 · SurroundOcc generation | mIoU: 12.04 -> 11.45 | Hurt |
| lower KL weight from 0.00001 to 0.000001 (ablation) | Doe-2 · SurroundOcc reconstruction | mIoU: 19.65 -> 19.73 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- OmniNWM — OmniNWM: Unifying the State-Action-Reward Triad for Closed-Loop Panoramic Driving Navigation World Models (ECCV 2026) · better on 2 of 2 shared results
- M2Depth — M^2Depth: Self-supervised Two-Frame Multi-camera Metric Depth Estimation (ECCV 2024) · better on 2 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
- GaussianWorld: Gaussian World Model for Streaming 3D Occupancy Prediction (CVPR 2025 · similarity 0.89)
- Semi-Supervised Vision-Centric 3D Occupancy World Model for Autonomous Driving (ICLR 2025 · similarity 0.89)
- Driving into the Future: Multiview Visual Forecasting and Planning with World Model for Autonomous Driving (CVPR 2024 · similarity 0.88)
- COME: Adding Scene-Centric Forecasting Control to Occupancy World Model (NeurIPS 2025 · similarity 0.88)
- X-Scene: Large-Scale Driving Scene Generation with High Fidelity and Flexible Controllability (NeurIPS 2025 · similarity 0.88)
- Future-Aware End-to-End Driving: Bidirectional Modeling of Trajectory Planning and Scene Evolution (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.
