Research map / 3D vision / Depth and stereo
PolyLayout: Multi-room Manhattan Layout Estimation
Gustav Hanning, Shaohui Liu, Rémi Pautrat, Marc Pollefeys, Kalle Åström, Viktor Larsson
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 |
|---|---|---|---|
| share orientation across rooms (ablation) | PolyLayout · ASE test set (ablation) | 3D IoU: 93.5 -> 93.8 | Improved |
| also share floor and ceiling height (ablation) | PolyLayout · ASE test set (ablation) | 3D IoU: 93.8 -> 94.3 | Improved |
| use DINOv2 ViT-S/14 instead of ResNet-101 (ablation) | PolyLayout · ASE test set (ablation) | 3D IoU: 82.7 -> 94.3 | Improved |
| use DINOv2 instead of ResNet-101 with featuremetric cost only (ablation) | PolyLayout (E_feat) · ASE test set (ablation) | 3D IoU: 24.6 -> 45.1 | Improved |
| add perimeter cost E_per (ablation) | PolyLayout · ASE test set (ablation) | 3D IoU: 93.5 -> 94.3 | Improved |
| add perimeter cost E_per (ablation) | PolyLayout · ASE test set (ablation) | Wall recall: 89.3 -> 89.0 | Hurt |
| concave hull initialization instead of cuboid (ablation) | PolyLayout · ASE test set (ablation) | 3D IoU: 90.5 -> 94.3 | Improved |
| concave hull initialization instead of circle (ablation) | PolyLayout · ASE test set (ablation) | 3D IoU: 93.2 -> 94.3 | Improved |
| add wall splitting (ablation) | PolyLayout · ASE test set (ablation) | 3D IoU: 93.6 -> 94.3 | Improved |
| add wall splitting (ablation) | PolyLayout · ASE test set (ablation) | Room recall: 82.2 -> 79.0 | Hurt |
| add polygon simplification (ablation) | PolyLayout · ASE test set (ablation) | 3D IoU: 92.9 -> 94.3 | Improved |
| add edge cost E_edge to featuremetric cost (ablation) | PolyLayout · ScanNet++ v2 cuboid rooms (point-based fitting) | 3D IoU: 63.5 -> 92.4 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- PSMNet — PSMNet: Position-Aware Stereo Merging Network for Room Layout Estimation (CVPR 2022) · better on 1 of 1 shared result
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
- GeoLayout: Geometry Driven Room Layout Estimation Based on Depth Maps of Planes (ECCV 2020 · similarity 0.91)
- HouseLayout3D: A Benchmark and Training-free Baseline for 3D Layout Estimation in the Wild (NeurIPS 2025 · similarity 0.91)
- LayoutNet: Reconstructing the 3D Room Layout From a Single RGB Image (CVPR 2018 · similarity 0.91)
- No More Ambiguity in 360deg Room Layout via Bi-Layout Estimation (CVPR 2024 · similarity 0.89)
- OptiScene: LLM-driven Indoor Scene Layout Generation via Scaled Human-aligned Data Synthesis and Multi-Stage Preference Optimization (NeurIPS 2025 · similarity 0.89)
- Unposed Sparse Views Room Layout Reconstruction in the Age of Pretrain Model (ICLR 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.
