ArcPro: Architectural Programs for Structured 3D Abstraction of Sparse Points
Qirui Huang, Runze Zhang, Kangjun Liu, Minglun Gong, Hao Zhang, Hui Huang
2025Year
5Top-tier citations
Abstract
Figure 1 . Structured 3D abstraction (bottom and top right for a zoom-in) in the form of architectural programs, obtained by our method ArcPro, which takes as input an extremely sparse point cloud (only ≈300 points) segment over each building. Despite such low-density, non-uniform, and noisy inputs, our method produces clean, low-face-count meshes that structurally conform to the real building objects. ArcPro can process 1,090 buildings over an area of 2.92 km 2 in approximately 37 seconds on a single 4090 GPU.
Ask about this paper
Your agent reads all of it.
Lune indexed this paper to the last equation, along with the top-tier papers that cite it. Ask a question and the answer quotes them.
Cited by top-tier papers5
- ImmerseGen: Agent-Guided Immersive World Generation with Alpha-Textured ProxiesJinyan Yuan, Bangbang Yang, Keke Wang, Panwang Pan et al.IEEE VR 2026 · 2 citations
- BuildAnyPoint: 3D Building Structured Abstraction from Diverse Point CloudsTongyan Hua, Haoran Gong, Yuan Liu, Di Wang et al.CVPR 2026 · 1 citation
- Fast SceneScript: Fast and Accurate Language‑Based 3D Scene Understanding via Multi‑Token PredictionRuihong Yin, Xuepeng Shi, Oleksandr Bailo, Marco Manfredi et al.CVPR 2026 · 1 citation
- BuildingWorld: A Structured 3D Building Dataset for Urban Foundation ModelsShangfeng Huang, Ruisheng Wang, Xin WangAAAI 2026
- BuildingGPT: Auto-Regressive Building Wireframe Reconstruction Model with Reinforcement LearningYuzhou Liu, Lingjie Zhu, Hanqiao Ye, Yujun Liu et al.CVPR 2026
Builds on6
- Unsupervised learning for cuboid shape abstraction via joint segmentation from point cloudsKaizhi Yang, Xuejin ChenSIGGRAPH 2021 · 52 citations
- UrbanBIS: a Large-scale Benchmark for Fine-grained Urban Building Instance SegmentationGuoqing Yang, Fuyou Xue, Qi Zhang, Ke Xie et al.SIGGRAPH 2023 · 39 citations
- RIM-Net: Recursive Implicit Fields for Unsupervised Learning of Hierarchical Shape StructuresChengjie Niu, Manyi Li, Kai Xu, Hao ZhangCVPR 2022 · 18 citations
- DPF-Net: Combining Explicit Shape Priors in Deformable Primitive Field for Unsupervised Structural Reconstruction of 3D ObjectsQingyao Shuai, Chi Zhang, Kaizhi Yang, Xuejin ChenICCV 2023 · 14 citations
- Visual Programming: Compositional visual reasoning without trainingTanmay Gupta, Aniruddha KembhaviCVPR 2023
Related papers
- PC2WF: 3D Wireframe Reconstruction from Raw Point CloudsYujia Liu, Stefano D'Aronco, Konrad Schindler, Jan Dirk WegnerICLR 2021 · 60 citations
- MonteFloor: Extending MCTS for Reconstructing Accurate Large-Scale Floor PlansSinisa Stekovic, Mahdi Rad, Friedrich Fraundorfer, Vincent LepetitICCV 2021 · 43 citations
- Low-poly Mesh Generation for Building ModelsXifeng Gao, Kui Wu, Zherong PanSIGGRAPH 2022 · 23 citations
- BuildingNet: Learning to Label 3D BuildingsPratheba Selvaraju, Mohamed Nabail, Marios Loizou, Maria Maslioukova et al.ICCV 2021 · 54 citations
- Cost Volume Pyramid Based Depth Inference for Multi-View StereoJiayu Yang, Wei Mao, José M. Álvarez, Miaomiao LiuCVPR 2020
