Research map / 3D vision / 3D generation
PatchScene: Patch-based Voxel Diffusion Model for Large-Scale Scene Completion
Qingdong Xu, Jiajun Zhu, Shilin Zhu, Xinjing He, Chao Lu, Huanran Wang, Jiyao Zhang
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 temporal fusion across frames (ablation) | PatchScene · SemanticKITTI | RMSE t0 to t1: 0.155 -> 0.086 | Improved |
| add temporal fusion across frames (ablation) | PatchScene · SemanticKITTI | JSD BEV: 0.371 -> 0.372 | No clear effect |
| use random coupling in overlap regions instead of no fusion (ablation) | PatchScene · SemanticKITTI | Chamfer Distance: 0.348 -> 0.319 | Improved |
| use random coupling instead of average addition (ablation) | PatchScene · SemanticKITTI | Chamfer Distance: 0.351 -> 0.319 | Improved |
| use random coupling instead of average addition (ablation) | PatchScene · SemanticKITTI | JSD 3D: 0.439 -> 0.444 | Mixed |
| annular outward generation instead of linear order (ablation) | PatchScene · SemanticKITTI | Chamfer Distance: 0.451 -> 0.319 | Improved |
| annular outward instead of annular inward (ablation) | PatchScene · SemanticKITTI | Chamfer Distance: 0.391 -> 0.319 | Improved |
| increase denoising timesteps from 10 to 50 (ablation) | PatchScene · SemanticKITTI | Chamfer Distance: 0.319 -> 0.357 | Hurt |
| reduce denoising timesteps from 10 to 5 (ablation) | PatchScene · SemanticKITTI | Chamfer Distance: 0.319 -> 0.310 | Mixed |
| patch-based voxel diffusion vs prior refined methods | PatchScene · SemanticKITTI | Voxel IoU 0.5: 36.6 -> 45.3 | Improved |
Closest papers by meaning
- Scaling Diffusion Models to Real-World 3D LiDAR Scene Completion (CVPR 2024 · similarity 0.91)
- Towards Realistic Scene Generation with LiDAR Diffusion Models (CVPR 2024 · similarity 0.89)
- Spiral: Semantic-Aware Progressive LiDAR Scene Generation and Understanding (NeurIPS 2025 · similarity 0.88)
- H2GFormer: Horizontal-to-Global Voxel Transformer for 3D Semantic Scene Completion (AAAI 2024 · similarity 0.87)
- Outdoor Scene Extrapolation with Hierarchical Generative Cellular Automata (CVPR 2024 · similarity 0.87)
- GeoComplete: Geometry-Aware Diffusion for Reference-Driven Image Completion (NeurIPS 2025 · similarity 0.87)
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.
