Research map / 3D vision / 3D generation
LaGen: Towards Autoregressive LiDAR Scene Generation
Sizhuo Zhou, Xiaosong Jia, Fanrui Zhang, Junjie Li, Juyong Zhang, Yukang Feng, Jianwen Sun, Songbur Wong, Junqi You, Junchi Yan
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 scene decoupling estimation (SDE) (ablation) | LaGen · Temporal LiDAR generation | CTC (step 4): 2.49 -> 2.22 | Improved |
| add scene decoupling estimation (SDE) (ablation) | LaGen · LiDAR generation | Chamfer Distance at 8.0s: 2.23 -> 2.16 | Improved |
| add noise modulation (NM) (ablation) | LaGen · Temporal LiDAR generation | CTC (step 4): 2.77 -> 2.22 | Improved |
| add noise modulation (NM) (ablation) | LaGen · LiDAR generation | Chamfer Distance at 9.0s: 3.16 -> 2.25 | Improved |
| use 50 denoising steps instead of 10 (ablation) | LaGen · LiDAR generation | Chamfer Distance at 9.5s: 2.60 -> 2.34 | Improved |
| use 50 steps instead of 10 (ablation) | LaGen · LiDAR generation | time per frame: 0.21 -> 0.43 s/frame | Hurt |
| raise denoising steps from 50 to 200 (ablation) | LaGen · LiDAR generation | Chamfer Distance at 9.5s: 2.34 -> 2.20 | No clear effect |
| add bounding box conditioning to LaGen | LaGen · LiDAR prediction (9.5s) | Chamfer Distance: 2.78 -> 2.34 | Improved |
| randomly shift bounding boxes by up to 8 m at inference (ablation) | LaGen · nuScenes LiDAR generation | MMD: 0.109 -> 0.335 | Hurt |
| randomly drop 50% of bounding boxes at inference (ablation) | LaGen · nuScenes LiDAR generation | MMD: 0.109 -> 0.463 | Hurt |
| LaGen vs LiDARCrafter on nuScenes | LaGen · nuScenes LiDAR generation | MMD: 1.52 -> 0.11 | Improved |
| LaGen vs LiDARCrafter on TTCE | LaGen · Temporal LiDAR generation | TTCE (step 4): 3.56 vs 3.62 | No clear effect |
Closest papers by meaning
- Genesis: Multimodal Driving Scene Generation with Spatio-Temporal and Cross-Modal Consistency (NeurIPS 2025 · similarity 0.89)
- Outdoor Scene Extrapolation with Hierarchical Generative Cellular Automata (CVPR 2024 · similarity 0.88)
- GenAssets: Generating in-the-wild 3D Assets in Latent Space (CVPR 2025 · similarity 0.88)
- UniScene: Unified Occupancy-centric Driving Scene Generation (CVPR 2025 · similarity 0.88)
- Towards Realistic Scene Generation with LiDAR Diffusion Models (CVPR 2024 · similarity 0.88)
- OLiDM: Object-aware LiDAR Diffusion Models for Autonomous Driving (AAAI 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.
