Research map / 3D vision / Depth and stereo
Quasi-Globally Optimal and Efficient Vanishing Point Estimation in Manhattan World
Haoang Li, Ji Zhao, Jean-Charles Bazin, Wen Chen, Zhe Liu, Yun-Hui Liu
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 |
|---|---|---|---|
| hybridize sampling of two lines with branch-and-bound search of the third DOF, instead of pure BnB over full rotation space | Manhattan frame rotation estimation · Synthetic (2 VPs, 100 inliers, sigma=4) | iterations: 318 -> 9 | Improved |
| hybridize sampling with BnB instead of pure BnB | Manhattan frame rotation estimation · Synthetic (2 VPs, 100 inliers, sigma=4) | computational time (s): 3.501 -> 0.276 | Improved |
| search the third DOF after sampling two lines, instead of sampling three lines as in RANSAC | RANSAC-style sampling · Real image with LSD lines (Fig. 13) | true inliers: 59 -> 69 | Improved |
| use sampling to restrict the search to the true Manhattan frame instead of global inlier maximization | BnB · Real image with LSD lines, many outliers | clustering correctness: 81 fake inliers (BnB) vs 69 true inliers | Improved |
| give RANSAC and J-Linkage more time budget, up to the budget of our method | RANSAC, J-Linkage · Synthetic, 40% outliers | F1-score: time budget from t/20 to t | Improved |
| increase outlier ratio above 30% | RANSAC and J-Linkage · Synthetic, 150 lines, sigma=3 | precision and recall: more than 30% outliers | Hurt |
Closest papers by meaning
- Quasi-Globally Optimal and Near/True Real-Time Vanishing Point Estimation in Manhattan World (TPAMI 2020 · similarity 0.98)
- Globally Optimal and Efficient Vanishing Point Estimation in Atlanta World (ECCV 2020 · similarity 0.92)
- Convex Relaxation for Robust Vanishing Point Estimation in Manhattan World (CVPR 2025 · similarity 0.91)
- Vanishing Point Estimation in Uncalibrated Images with Prior Gravity Direction (ICCV 2023 · similarity 0.90)
- Deep Single Image Camera Calibration by Heatmap Regression to Recover Fisheye Images Under Manhattan World Assumption (CVPR 2024 · similarity 0.87)
- VPDETR: End-to-End Vanishing Point DEtection TRansformers (AAAI 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.
