Research map / 3D vision / 3D generation
Disentangling Orthogonal Planes for Indoor Panoramic Room Layout Estimation With Cross-Scale Distortion Awareness
Zhijie Shen, Zishuo Zheng, Chunyu Lin, Lang Nie, Kang Liao, Shuai Zheng, Yao Zhao
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 DOPNet instead of LGT-Net | Cuboid rooms trained on PanoContext + whole Stanford 2D-3D · PanoContext + Stanford 2D-3D | 3DIoU (%): 85.16 -> 85.46 | Improved |
| Use DOPNet instead of Dula-Netv2 | Cuboid rooms trained on Stanford 2D-3D + whole PanoContext · Stanford 2D-3D + PanoContext | 3DIoU (%): 86.60 -> 85.47 | Hurt |
| Use DOPNet instead of LGT-Net | DOPNet · MatterportLayout | 3DIoU (%): 81.11 -> 81.70 | Improved |
| Use DOPNet instead of LGT-Net | DOPNet · ZInD | 3DIoU (%): 89.95 -> 90.13 | Improved |
| Add cross-scale interaction (ablation) | DOPNet feature assembling · MatterportLayout | 3DIoU (%): 80.71 -> 81.34 | Improved |
| Add distortion elimination and feature assembling (ablation) | DOPNet feature extraction · MatterportLayout | 3DIoU (%): 80.10 -> 81.34 | Improved |
| Add orthogonal plane disentanglement (ablation) | DOPNet · MatterportLayout | 3DIoU (%): 80.60 -> 81.34 | Improved |
| Add soft-flipping fusion (ablation) | DOPNet · MatterportLayout | 3DIoU (%): 80.35 -> 81.34 | Improved |
| Add discriminative channel generation (ablation) | DOPNet sequence reconstruction · MatterportLayout | 3DIoU (%): 80.23 -> 81.34 | Improved |
| Add self-attention for long-range dependencies (ablation) | DOPNet sequence reconstruction · MatterportLayout | 3DIoU (%): 80.97 -> 81.34 | Improved |
| Add cross-attention residuals (ablation) | DOPNet sequence reconstruction · MatterportLayout | 3DIoU (%): 80.45 -> 81.34 | 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.
- LGT-Net — LGT-Net: Indoor Panoramic Room Layout Estimation With Geometry-Aware Transformer Network (CVPR 2022) · better on 3 of 3 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
- 360 Layout Estimation via Orthogonal Planes Disentanglement and Multi-View Geometric Consistency Perception (TPAMI 2024 · similarity 0.92)
- LGT-Net: Indoor Panoramic Room Layout Estimation With Geometry-Aware Transformer Network (CVPR 2022 · similarity 0.89)
- HorizonNet: Learning Room Layout With 1D Representation and Pano Stretch Data Augmentation (CVPR 2019 · similarity 0.88)
- GeoLayout: Geometry Driven Room Layout Estimation Based on Depth Maps of Planes (ECCV 2020 · similarity 0.88)
- HouseLayout3D: A Benchmark and Training-free Baseline for 3D Layout Estimation in the Wild (NeurIPS 2025 · similarity 0.87)
- No More Ambiguity in 360deg Room Layout via Bi-Layout Estimation (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.
