Research map / 3D vision / 3D generation
360-MLC: Multi-view Layout Consistency for Self-training and Hyper-parameter Tuning
Bolivar Enrique Solarte, Chin-Hsuan Wu, Yueh-Cheng Liu, Yi-Hsuan Tsai, Min Sun
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 multi-view consistency self-training rather than SSLayout360* | HorizonNet with MatterportLayout pre-training, Setting 1 · MP3D-FPE | 2D IoU (%): 70.53 -> 71.50 | Improved |
| Use multi-view consistency self-training rather than SSLayout360* | HorizonNet with ZInD pre-training, Setting 1 · MP3D-FPE | 2D IoU (%): 62.62 -> 65.60 | Improved |
| Use multi-view consistency self-training rather than SSLayout360* | HorizonNet with LayoutNet pre-training, Setting 1 · MP3D-FPE | 2D IoU (%): 67.48 -> 69.40 | Improved |
| Use multi-view consistency self-training rather than SSLayout360-ST* | HorizonNet with MatterportLayout pre-training, Setting 2 · MP3D-FPE | 2D IoU (%): 70.58 -> 71.50 | Improved |
| Use multi-view consistency self-training rather than SSLayout360-ST* | HorizonNet with ZInD pre-training, Setting 2 · MP3D-FPE | 2D IoU (%): 56.52 -> 64.45 | Improved |
| Use multi-view consistency self-training rather than SSLayout360-ST* | HorizonNet with LayoutNet pre-training, Setting 2 · MP3D-FPE | 2D IoU (%): 66.14 -> 69.12 | Improved |
| Increase pseudo-label source frames from 10% to 50% (ablation) | 360-MLC with ZInD pre-training · MP3D-FPE, Setting 2 | 2D IoU (%): 57.29 -> 59.55 | Improved |
| Increase pseudo-label source frames from 50% to 100% (ablation) | 360-MLC with ZInD pre-training · MP3D-FPE, Setting 2 | 2D IoU (%): 59.55 -> 64.45 | Improved |
| Use median rather than mean to generate pseudo-labels (ablation) | 360-MLC with ZInD pre-training · MP3D-FPE, Setting 2 | 2D IoU (%): 46.17 -> 64.45 | Improved |
| Add uncertainty weighting to the boundary consistency loss (ablation) | 360-MLC with ZInD pre-training · MP3D-FPE, Setting 2 | 2D IoU (%): 62.81 -> 64.45 | Improved |
| Use estimated rather than ground-truth camera poses (ablation) | 360-MLC with MatterportLayout pre-training · MP3D-FPE | 2D IoU (%): 71.50 -> 70.85 | Hurt |
| Use noisy rather than ground-truth camera poses (ablation) | 360-MLC with MatterportLayout pre-training · MP3D-FPE | 2D IoU (%): 71.50 -> 66.05 | Hurt |
Later papers that compared with this method
2 later accepted papers report 360-MLC as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- Self-Training Room Layout via Geometry-aware Ray-casting (ECCV 2024 · better on 8 of 8 shared results)
- 360 Layout Estimation via Orthogonal Planes Disentanglement and Multi-View Geometric Consistency Perception (TPAMI 2024 · better on 2 of 2 shared results)
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.
- SSLayout360* — SSLayout360: Semi-Supervised Indoor Layout Estimation From 360deg Panorama (CVPR 2021) · 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
- No More Ambiguity in 360deg Room Layout via Bi-Layout Estimation (CVPR 2024 · similarity 0.89)
- HouseLayout3D: A Benchmark and Training-free Baseline for 3D Layout Estimation in the Wild (NeurIPS 2025 · similarity 0.87)
- Unposed Sparse Views Room Layout Reconstruction in the Age of Pretrain Model (ICLR 2025 · similarity 0.86)
- OptiScene: LLM-driven Indoor Scene Layout Generation via Scaled Human-aligned Data Synthesis and Multi-Stage Preference Optimization (NeurIPS 2025 · similarity 0.85)
- MvFFN: Multi-view Floor-Plan Feed-Forward Network for Unposed Wide-Baseline Panorama Layout Reconstruction (NeurIPS 2026 · similarity 0.85)
- Direct Numerical Layout Generation for 3D Indoor Scene Synthesis via Spatial Reasoning (NeurIPS 2025 · similarity 0.85)
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.
