Research map / 3D vision / 3D generation
Occ3D: A Large-Scale 3D Occupancy Prediction Benchmark for Autonomous Driving
Xiaoyu Tian, Tao Jiang, Longfei Yun, Yucheng Mao, Huitong Yang, Yue Wang, Yilun Wang, Hang Zhao
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 |
|---|---|---|---|
| replace BEVFormer with CTF-Occ coarse-to-fine encoder | BEVFormer · Occ3D-nuScenes | mIoU: 26.88 -> 28.53 | Improved |
| replace BEVFormer with CTF-Occ coarse-to-fine encoder | BEVFormer · Occ3D-Waymo | mIoU: 16.76 -> 18.73 | Improved |
| add OHEM loss to top-k token selection (ablation) | CTF-Occ · Occ3D-Waymo | mIoU: 14.06 -> 18.43 | Improved |
| use top-k instead of random token selection (ablation) | CTF-Occ · Occ3D-Waymo | mIoU: 16.62 -> 18.43 | Improved |
| use top-k instead of uncertain token selection (ablation) | CTF-Occ · Occ3D-Waymo | mIoU: 17.37 -> 18.43 | No clear effect |
| add OHEM loss (small objects) (ablation) | CTF-Occ · Occ3D-Waymo pedestrian | IoU: 4.16 -> 7.04 | Improved |
| add multi-frame aggregation to single-frame points (ablation) | label pipeline · 3D-2D consistency | mIoU: 13.32 -> 17.65 | Improved |
| add voxelization (size 0.1) to multi-frame points (ablation) | label pipeline · 3D-2D consistency | mIoU: 17.65 -> 41.17 | Improved |
| add mesh reconstruction at voxel 0.1 (ablation) | label pipeline · 3D-2D consistency | mIoU: 41.17 -> 41.99 | Improved |
| reduce voxel size from 0.1 to 0.05 (with mesh) (ablation) | label pipeline · 3D-2D consistency | mIoU: 41.99 -> 43.58 | Improved |
| add image-guided voxel refinement (ablation) | label pipeline · 3D-2D consistency | mIoU: 43.58 -> 58.50 | Improved |
| add LiDAR input to BEVFormer (fusion) | BEVFormer · Occ3D-Waymo | mIoU: 16.76 -> 39.05 | Improved |
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.
- OccFormer — OccFormer: Dual-path Transformer for Vision-based 3D Semantic Occupancy Prediction (ICCV 2023) · better on 1 of 1 shared result
- MonoScene — MonoScene: Monocular 3D Semantic Scene Completion (CVPR 2022) · better on 1 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
- Achieving Speed-Accuracy Balance in Vision-based 3D Occupancy Prediction via Geometric-Semantic Disentanglement (AAAI 2025 · similarity 0.93)
- OctOcc: High-Resolution 3D Occupancy Prediction with Octree (AAAI 2024 · similarity 0.92)
- CSV-Occ: Fusing Multi-frame Alignment for Occupancy Prediction with Temporal Cross State Space Model and Central Voting Mechanism (ICML 2025 · similarity 0.91)
- Semi-Supervised Vision-Centric 3D Occupancy World Model for Autonomous Driving (ICLR 2025 · similarity 0.91)
- Cam4DOcc: Benchmark for Camera-Only 4D Occupancy Forecasting in Autonomous Driving Applications (CVPR 2024 · similarity 0.90)
- COTR: Compact Occupancy TRansformer for Vision-based 3D Occupancy Prediction (CVPR 2024 · similarity 0.90)
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.
