Research map / Detection and segmentation / CNN features
Supercharging Floorplan Localization with Semantic Rays
Yuval Grader, Hadar Averbuch-Elor
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 coarse-to-fine refinement (ablation) | Base model · S3D floorplan localization | R@1m 30deg: 48.44 -> 52.61 | Improved |
| add room type predictions (ablation) | Base model · S3D floorplan localization | R@1m 30deg: 48.44 -> 54.04 | Improved |
| add both room predictions and refinement (ablation) | Base model · S3D floorplan localization | R@1m 30deg: 48.44 -> 57.49 | Improved |
| add majority-vote semantic interpolation instead of linear interpolation (ablation) | Base model · S3D floorplan localization | R@1m 30deg: 47.99 -> 48.44 | Improved |
| add our refinement module to F3Loc | F3Loc · S3D floorplan localization | R@1m 30deg: 21.3 -> 29.59 | Improved |
| add our refinement module to F3Loc | F3Loc · ZInD floorplan localization | R@1m 30deg: 11.46 -> 14.21 | Improved |
| use external masking (ablation) | Ours (room labels) · S3D floorplan localization | R@1m 30deg: 57.61 -> 57.82 | No clear effect |
| fuse semantic rays with depth rays instead of depth only (w_d=1.0) (ablation) | structural-semantic probability volume · S3D floorplan localization | R@1m 30deg: 29.05 -> 49.48 | Improved |
| use semantic rays only (w_s=1.0) instead of fused (ablation) | structural-semantic probability volume · S3D floorplan localization | R@1m 30deg: 49.48 -> 7.27 | Hurt |
| increase Top-K refined candidates from 1 to 5 (ablation) | Ours (no room labels) · S3D floorplan localization | R@1m 30deg: 48.59 -> 52.61 | Improved |
| lower refinement threshold delta_res to 0.05m from 1 (ablation) | Ours · S3D floorplan localization | R@0.1m: 7.85 -> 18.40 | Improved |
| raise refinement threshold delta_res to 1m from 0.05m (ablation) | Ours · S3D floorplan localization | R@1m 30deg: 70.08 -> 76.65 | Improved |
Closest papers by meaning
- Fusion of Depth and Semantics for Probabilistic Floorplan Localization (CVPR 2026 · similarity 0.93)
- Perspective from a Broader Context: Can Room Style Knowledge Help Visual Floorplan Localization? (AAAI 2026 · similarity 0.90)
- F3Loc: Fusion and Filtering for Floorplan Localization (CVPR 2024 · similarity 0.88)
- UnLoc: Leveraging Depth Uncertainties for Floorplan Localization (ICLR 2026 · similarity 0.88)
- HouseLayout3D: A Benchmark and Training-free Baseline for 3D Layout Estimation in the Wild (NeurIPS 2025 · similarity 0.86)
- GSDiff: Synthesizing Vector Floorplans via Geometry-enhanced Structural Graph Generation (AAAI 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.
