Research map / Point clouds / Segmentation and detection
Building3D: A Urban-Scale Dataset and Benchmarks for Learning Roof Structures from Point Clouds
Ruisheng Wang, Shangfeng Huang, Hongxin Yang
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 |
|---|---|---|---|
| use Point Transformer based supervised pipeline instead of Point2Roof | Building3D roof wireframe reconstruction · Building3D Tallinn | edge F1: 0.14 -> 0.36 | Improved |
| increase labeled data ratio for fine-tuning self-supervised model (ablation) | Our self-supervised · Building3D Tallinn | ACO: 0.57 -> 0.28 | Improved |
| use linear self-attention variant of Point-MAE | Point-MAE · Building3D Tallinn (50% labels) | ACO: 0.33 -> 0.30 | Improved |
| use linear self-attention variant of Point-MAE | Point-MAE · Building3D entry-level | edge F1: 0.35 -> 0.52 | Improved |
| use Point Transformer plus graph edge generation instead of FG-Net | supervised baselines · Building3D entry-level | edge F1: 0.52 -> 0.61 | Improved |
| deep learning methods vs recall of edges | all deep baselines · Building3D | edge recall | Hurt |
| use deep learning method instead of handcrafted methods | Our supervised method vs PolyFit · Building A | RMSE: 0.080 -> 0.490 | Hurt |
| use deep learning method instead of handcrafted methods | Our method vs PolyFit · Building A | faces: 31 -> 19 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- PolyFit — PolyFit: Polygonal Surface Reconstruction From Point Clouds (ICCV 2017) · better on 1 of 2 shared results
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
- BuildingWorld: A Structured 3D Building Dataset for Urban Foundation Models (AAAI 2026 · similarity 0.92)
- USCILab3D: A Large-scale, Long-term, Semantically Annotated Outdoor Dataset (NeurIPS 2024 · similarity 0.87)
- LiDAR-Net: A Real-scanned 3D Point Cloud Dataset for Indoor Scenes (CVPR 2024 · similarity 0.86)
- BWFormer: Building Wireframe Reconstruction from Airborne LiDAR Point Cloud with Transformer (CVPR 2025 · similarity 0.86)
- 3D Building Reconstruction from Monocular Remote Sensing Images with Multi-level Supervisions (CVPR 2024 · similarity 0.86)
- PBWR: Parametric-Building-Wireframe Reconstruction from Aerial LiDAR Point Clouds (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.
