Research map / 3D vision / 3D generation
SparseWorld-TC: Trajectory-Conditioned Sparse Occupancy World Model
Jiayuan Du, Yiming Zhao, Zhenglong Guo, Yong Pan, Wenbo Hou, Zhihui Hao, Kun Zhan, Qijun Chen
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 predicted trajectory conditioning (ablation) | SparseWorld-TC-Small · Occ3D-nuScenes 1-3s | avg mIoU (%): 15.44 -> 21.57 | Improved |
| use ground-truth trajectory instead of predicted (ablation) | SparseWorld-TC-Small · Occ3D-nuScenes 1-3s | avg mIoU (%): 21.57 -> 25.60 | Improved |
| add predicted trajectory conditioning (ablation) | SparseWorld-TC-Small · Occ3D-nuScenes 1-3s | avg IoU (%): 32.19 -> 44.76 | Improved |
| random-ensemble horizon supervision instead of fixed horizon (ablation) | SparseWorld-TC-Large · Occ3D-nuScenes 1-8s | avg mIoU (%): 20.36 -> 22.33 | Improved |
| random-ensemble horizon supervision instead of fixed horizon (ablation) | SparseWorld-TC-Large · Occ3D-nuScenes 1-8s | avg IoU (%): 43.25 -> 45.35 | Improved |
| swap ResNet-50 for DINOv3-Base backbone (ablation) | SparseWorld-TC-Large · Occ3D-nuScenes 1-3s | avg mIoU (%): 26.42 -> 29.89 | Improved |
| increase anchors from 600x128 to 4800x16 (Small to Large) (ablation) | SparseWorld-TC · Occ3D-nuScenes 1-3s | avg mIoU (%) / FPS: mIoU 25.60 -> 26.42; FPS 9.35 -> 3.58 | Mixed |
| sparse-anchor transformer instead of VAE/BEV pipeline (vs COME) | SparseWorld-TC-Large · Occ3D-nuScenes 1-3s | avg mIoU (%): 22.26 -> 26.42 | Improved |
| camera-only sparse model vs GT-occupancy-input COME on long horizon | SparseWorld-TC-Large · Occ3D-nuScenes 1-8s | avg IoU (%): 29.96 -> 45.35 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 3 on most shared results.
- COME — COME: Adding Scene-Centric Forecasting Control to Occupancy World Model (NeurIPS 2025) · better on 2 of 2 shared results
- SparseWorld — SparseWorld: A Flexible, Adaptive, and Efficient 4D Occupancy World Model Powered by Sparse and Dynamic Queries (AAAI 2026) · better on 1 of 1 shared result
- I2-World — I2-World: Intra-Inter Tokenization for Efficient Dynamic 4D Scene Forecasting (ICCV 2025) · better on 0 of 1 shared result
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
- STCOcc: Sparse Spatial-Temporal Cascade Renovation for 3D Occupancy and Scene Flow Prediction (CVPR 2025 · similarity 0.89)
- ODG: Occupancy Prediction Using Dual Gaussians (NeurIPS 2025 · similarity 0.88)
- Spatiotemporal Decoupling for Efficient Vision-Based Occupancy Forecasting (CVPR 2025 · similarity 0.88)
- Semi-Supervised Vision-Centric 3D Occupancy World Model for Autonomous Driving (ICLR 2025 · similarity 0.88)
- GaussianWorld: Gaussian World Model for Streaming 3D Occupancy Prediction (CVPR 2025 · similarity 0.87)
- SparseOcc: Rethinking Sparse Latent Representation for Vision-Based Semantic Occupancy Prediction (CVPR 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.
