Research map / 3D vision / 3D generation
Scene as Occupancy
Wenwen Tong, Chonghao Sima, Tai Wang, Li Chen, Silei Wu, Hanming Deng, Yi Gu, Lewei Lu, Ping Luo, Dahua Lin, Hongyang Li
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 |
|---|---|---|---|
| decode voxel features from BEV features with a cascade voxel decoder instead of BEVNet (ablation) | OccNet · OpenOcc val (ResNet50) | mIoU: 17.37 -> 19.48 | Improved |
| use cascade voxel decoder instead of direct VoxelNet (ablation) | OccNet · OpenOcc val (ResNet50) | mIoU: 19.06 -> 19.48; FPS 1.9 -> 2.6; memory 23G -> 18G | Improved |
| add occupancy-based scene completion features as perception input for planning rasterization instead of bounding boxes | ST-P3-style planner · nuScenes planning | Collision rate 1s (%): 0.27 -> 0.21 | Improved |
| use all 16 occupancy classes for planning (ablation) | OccNet planning · nuScenes planning | L2 3s (m): L2 3s 3.07 -> 2.99; collision 2s 0.55 -> 0.59 | Mixed |
| jointly train scene completion with 3D detection (ablation) | BEVNet, VoxNet, OccNet · nuScenes val 3D detection | mAP / NDS: BEVNet mAP 0.259 -> 0.271; OccNet NDS 0.382 -> 0.390 | Improved |
| joint training with occupancy slightly worsens box translation and scale errors (ablation) | BEVNet · nuScenes val 3D detection | mATE: 0.828 -> 0.835 | Hurt |
| pretrain on occupancy instead of detection for BEV segmentation fine-tuning | BEV segmentation · nuScenes BEV segmentation | IoU main value: 18.18 -> 19.17 | Improved |
| pretrain on occupancy rather than detection for planning | ST-P3 planner · nuScenes planning | Collision rate 3s (%): 0.82 -> 1.03 | Hurt |
| use finer voxel size 0.25m instead of 0.5m for irregular objects (ablation) | Occupancy vs detection · nuScenes irregular objects | mIoU_ir gap: +2.21 -> +3.51 | Improved |
| increase history frames for temporal self-attention from 2 to 4 (ablation) | OccNet · OpenOcc val (ResNet50) | mIoU: 19.48 -> 20.81 | Improved |
| use only one history frame (ablation) | OccNet · OpenOcc val (ResNet50) | mIoU: 19.21 (0 frames) -> 18.35 (1 frame) | Hurt |
| decrease voxel size for LiDAR segmentation transfer (ablation) | OccNet · nuScenes val LiDAR segmentation | mIoU: 46.60 (1.0m) -> 47.29 (0.5m) -> 53.00 (0.25m) | 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.
- BEVDepth — BEVDepth: Acquisition of Reliable Depth for Multi-View 3D Object Detection (AAAI 2023) · 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
- Towards Flexible 3D Perception: Object-Centric Occupancy Completion Augments 3D Object Detection (NeurIPS 2024 · similarity 0.93)
- Achieving Speed-Accuracy Balance in Vision-based 3D Occupancy Prediction via Geometric-Semantic Disentanglement (AAAI 2025 · 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 0.92)
- OctOcc: High-Resolution 3D Occupancy Prediction with Octree (AAAI 2024 · similarity 0.92)
- PanoOcc: Unified Occupancy Representation for Camera-based 3D Panoptic Segmentation (CVPR 2024 · similarity 0.91)
- COTR: Compact Occupancy TRansformer for Vision-based 3D Occupancy Prediction (CVPR 2024 · similarity 0.91)
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.
