Research map / 3D vision / Depth and stereo
Robust and Globally Optimal Manhattan Frame Estimation in Near Real Time
Kyungdon Joo, Tae-Hyun Oh, Junsik Kim, In So Kweon
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 EGI-based relaxed bound instead of exhaustive bound computation | branch-and-bound MF estimation · synthetic data | bound computation time: 108.178s -> 4.6ms | Improved |
| use EGI-based relaxed bound instead of exhaustive BnB | branch-and-bound MF estimation · NYUv2 | runtime: 117.06 -> 0.07 | Improved |
| relax the problem on the EGI instead of solving the exact problem | branch-and-bound MF estimation · NYUv2 | average angular error: 2.5 -> 2.63 | Hurt |
| use proposed BnB instead of RMFE | MF estimation · NYUv2 | average angular error: 3.27 -> 2.63 | Improved |
| use proposed BnB instead of MMF | MF estimation · NYUv2 | average angular error: 5.07 -> 2.63 | Improved |
| use proposed BnB instead of RTMF with IMU fusion | MF estimation · sequence dataset | angular RMSE: 3.05 -> 1.53 | Improved |
| use proposed BnB instead of vMF-RTMF alone | MF estimation · sequence dataset | angular RMSE: 6.39 -> 1.53 | Improved |
| increase EGI resolution parameter s (ablation) | proposed BnB · synthetic data | angular error and runtime | Mixed |
| vary inlier threshold tau (ablation) | proposed BnB · synthetic data | angular error | No clear effect |
| use proposed method instead of MMF for two MFs | multiple MF estimation · NYUv2 multiple MF | theta_x angular error: 13.8 -> 6.3 | Improved |
| use proposed method instead of MMF for one MF | multiple MF estimation · NYUv2 multiple MF | theta_x angular error: 5.2 -> 2.8 | Improved |
| use proposed method instead of RMFE for VP estimation | VP estimation · York Urban Dataset | horizon-error AUC: 77.62% -> 82.56% | Improved |
Closest papers by meaning
- Globally Optimal Manhattan Frame Estimation in Real-Time (CVPR 2016 · similarity 0.97)
- Globally Optimal and Efficient Manhattan Frame Estimation by Delimiting Rotation Search Space (ICCV 2021 · similarity 0.91)
- Quasi-Globally Optimal and Efficient Vanishing Point Estimation in Manhattan World (ICCV 2019 · similarity 0.89)
- Convex Relaxation for Robust Vanishing Point Estimation in Manhattan World (CVPR 2025 · similarity 0.85)
- Tensor-Based Synchronization and the Low-Rankness of the Block Trifocal Tensor (NeurIPS 2024 · similarity 0.83)
- Deep Single Image Camera Calibration by Heatmap Regression to Recover Fisheye Images Under Manhattan World Assumption (CVPR 2024 · similarity 0.83)
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.
