Research map / 3D vision / 3D generation
Scaling up Occupancy-Centric Driving Scene Generation: Dataset and Method
B R Li, Xin Jin, Hu Zhu, Hongsi Liu, Ruikai Li, Jiazhe Guo, Kaiwen Cai, Ma Ch, Yueming Jin, H Zhao, Xiaokang Yang, Wenjun Zeng
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 VAE 3D axial attention (ablation) | occupancy VAE · Nuplan-mini occupancy generation | mIoU: 21.42 -> 32.22 | Improved |
| add DiT temporal attention (ablation) | occupancy DiT · Nuplan-mini occupancy generation | mIoU: 19.32 -> 32.22 | Improved |
| add DiT spatial attention (ablation) | occupancy DiT · Nuplan-mini occupancy generation | mIoU: 15.67 -> 32.22 | Improved |
| add BEV condition (ablation) | occupancy generation · Nuplan-mini occupancy generation | mIoU: 17.13 -> 32.22 | Improved |
| add sparse rendered semantic map condition (ablation) | video generation · Nuplan-mini video generation | FID: 12.27 -> 8.32 | Improved |
| add sparse rendered depth map condition (ablation) | video generation · Nuplan-mini video generation | FID: 12.05 -> 8.32 | Improved |
| add unscented transform calibration (ablation) | Gaussian point map rendering · Nuplan-mini video generation | FID: 9.26 -> 8.32 | Improved |
| change Gaussian scale from 0.002 / 0.04 to 0.01 (ablation) | Gaussian point map rendering · Nuplan-mini video generation | FID: 8.76 / 9.21 -> 8.32 | Mixed |
| add sensor-specific embedding (ablation) | LiDAR generation · Nuplan-mini LiDAR generation | MMD (1e-5): 0.783 -> 0.457 | Improved |
| add histogram embedding (ablation) | LiDAR generation · Nuplan-mini LiDAR generation | MMD (1e-5): 0.997 -> 0.457 | Improved |
| add ray smoothness regularization (ablation) | LiDAR generation · Nuplan-mini LiDAR generation | MMD (1e-5): 0.694 -> 0.457 | Improved |
| scale generated training data from 10% to 100% of Nuplan-Occ | planning with UniAD · NAVSIM/Nuplan test | NC: 93.4 -> 95.7 | Improved |
Closest papers by meaning
- UniScene: Unified Occupancy-centric Driving Scene Generation (CVPR 2025 · similarity 0.94)
- SDGOCC: Semantic and Depth-Guided Bird's-Eye View Transformation for 3D Multimodal Occupancy Prediction (CVPR 2025 · similarity 0.91)
- Cam4DOcc: Benchmark for Camera-Only 4D Occupancy Forecasting in Autonomous Driving Applications (CVPR 2024 · similarity 0.90)
- Driving in the Occupancy World: Vision-Centric 4D Occupancy Forecasting and Planning via World Models for Autonomous Driving (AAAI 2025 · similarity 0.90)
- UnO: Unsupervised Occupancy Fields for Perception and Forecasting (CVPR 2024 · similarity 0.90)
- ODG: Occupancy Prediction Using Dual Gaussians (NeurIPS 2025 · 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.
