Research map / 3D vision / 3D generation
U4D: Uncertainty-Aware 4D World Modeling from LiDAR Sequences
Xiang Xu, Alan Liang, Youquan Liu, Linfeng Li, Lingdong Kong, Ziwei Liu, Qingshan Liu
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 entropy-based uncertainty region conditioning (ablation) | U4D · nuScenes (uncertainty region ablation) | FRD: 235.91 -> 223.96 | Improved |
| add random region conditioning instead of none (ablation) | U4D · nuScenes (uncertainty region ablation) | FRD: 235.91 -> 235.23 | No clear effect |
| use random region conditioning instead of none (ablation) | U4D · nuScenes (uncertainty region ablation) | ECE: 3.98 -> 4.35 | Hurt |
| add confidence-based region selection (ablation) | U4D · nuScenes (uncertainty region ablation) | FRD: 235.91 -> 228.24 | Improved |
| use entropy instead of confidence for region selection (ablation) | U4D · nuScenes (uncertainty region ablation) | ECE: 3.02 -> 2.72 | Improved |
| replace cascaded spatial-temporal with parallel addition fusion (ablation) | MoST block · nuScenes (MoST ablation) | FRD: 536.23 -> 242.81 | Improved |
| use concatenation instead of cascaded structure (ablation) | MoST block · nuScenes (MoST ablation) | FRD: 536.23 -> 242.43 | Improved |
| use adaptive fusion instead of concatenation (ablation) | MoST block · nuScenes (MoST ablation) | FRD: 242.43 -> 223.96 | Improved |
| use adaptive fusion instead of concatenation (ablation) | MoST block · nuScenes (MoST ablation) | FPD: 12.51 -> 12.90 | No clear effect |
| apply U4D generated data to semi-supervised segmentation | MinkUNet · nuScenes semi-supervised segmentation (1% labels) | mIoU: 58.3 -> 65.3 | Improved |
| apply U4D generated data to semi-supervised segmentation | MinkUNet · nuScenes semi-supervised segmentation (50% labels) | mIoU: 75.1 -> 76.4 | Improved |
| add U4D calibration | RangeNet++ · nuScenes val calibration | ECE (%): 4.57 -> 2.72 | Improved |
Closest papers by meaning
- MCD: Diverse Large-Scale Multi-Campus Dataset for Robot Perception (CVPR 2024 · similarity 0.87)
- 3D LiDAR Mapping in Dynamic Environments using a 4D Implicit Neural Representation (CVPR 2024 · similarity 0.87)
- Motion4D: Learning 3D-Consistent Motion and Semantics for 4D Scene Understanding (NeurIPS 2025 · similarity 0.86)
- Outdoor Scene Extrapolation with Hierarchical Generative Cellular Automata (CVPR 2024 · similarity 0.86)
- LiDAR4D: Dynamic Neural Fields for Novel Space-time View LiDAR Synthesis (CVPR 2024 · similarity 0.86)
- An Empirical Study of the Generalization Ability of Lidar 3D Object Detectors to Unseen Domains (CVPR 2024 · similarity 0.86)
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.
