Research map / 3D vision / Face and shape
MCTS With Refinement for Proposals Selection Games in Scene Understanding
Sinisa Stekovic, Mahdi Rad, Alireza Moradi, Friedrich Fraundorfer, Vincent Lepetit
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 |
|---|---|---|---|
| add refinement step (ablation) | MonteFloor · Structured3D | MA precision: 0.82 -> 0.90 | Improved |
| add refinement step (ablation) | MonteFloor · Structured3D | angle precision: 0.65 -> 0.86 | Improved |
| add angle loss L_ang (ablation) | MonteFloor · Structured3D | angle precision: 0.73 -> 0.86 | Improved |
| add global loss L_glob (ablation) | MonteFloor · Structured3D | room precision: 0.85 -> 0.96 | Improved |
| add L_0 term (ablation) | MonteFloor · Structured3D | angle precision: 0.84 -> 0.86 | Improved |
| add metric network f as selection score (ablation) | MonteFloor · Structured3D | MA recall: 0.17 -> 0.82 | Improved |
| add refinement step (ablation) | MonteRoom · Matterport3D-Layout validation | PE: 5.45 -> 4.56 | Improved |
| add L_2D term (ablation) | MonteRoom · Matterport3D-Layout validation | PE: 6.16 -> 4.56 | Improved |
| add L_3D term (ablation) | MonteRoom · Matterport3D-Layout validation | CE3D: 11.85 -> 11.47 | Improved |
| add regularization loss L_reg (ablation) | MonteRoom · Matterport3D-Layout validation | CE3D: 11.84 -> 11.47 | Improved |
| add MCTS selection and refinement over plane detection alone (ablation) | MonteRoom · Matterport3D-Layout validation | PE: 5.53 -> 4.56 | Improved |
| add refinement step (runtime cost) (ablation) | MonteFloor · Structured3D | computation time: 12 -> 71 seconds | Hurt |
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.
- GeoLayout — GeoLayout: Geometry Driven Room Layout Estimation Based on Depth Maps of Planes (ECCV 2020) · better on 5 of 5 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
- MonteFloor: Extending MCTS for Reconstructing Accurate Large-Scale Floor Plans (ICCV 2021 · similarity 0.95)
- Monte Carlo Scene Search for 3D Scene Understanding (CVPR 2021 · similarity 0.93)
- CASAGPT: Cuboid Arrangement and Scene Assembly for Interior Design (CVPR 2025 · similarity 0.86)
- PyTorchGeoNodes: Enabling Differentiable Shape Programs for 3D Shape Reconstruction (CVPR 2025 · similarity 0.85)
- HouseLayout3D: A Benchmark and Training-free Baseline for 3D Layout Estimation in the Wild (NeurIPS 2025 · similarity 0.85)
- PlanarGS: High-Fidelity Indoor 3D Gaussian Splatting Guided by Vision-Language Planar Priors (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.
