Research map / 3D vision / Depth and stereo
Estimation of Camera Locations in Highly Corrupted Scenarios: All About That Base, No Shape Trouble
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 IR-AAB edge filtering before LUD | LUD · Real-data datasets, excluding Ellis Island and Gendarmenmarkt | mean location error improvement rate: 38% | Helped |
| Use naive AAB edge filtering instead of no preprocessing | LUD · 12 real-data datasets, excluding Ellis Island and Gendarmenmarkt | mean location error improvement rate: 26.5% | Helped |
| Use 1DSfM edge filtering instead of no preprocessing | LUD · 12 real-data datasets, excluding Ellis Island and Gendarmenmarkt | mean location error improvement rate: 5.7% | Mixed |
| Add IR-AAB edge filtering before CLS | CLS · 12 real-data datasets, excluding Ellis Island and Gendarmenmarkt | mean and median location error improvement rate: more than 50% | Helped |
| Add naive AAB edge filtering before ShapeFit | ShapeFit · 12 real-data datasets, excluding Ellis Island and Gendarmenmarkt | mean location error improvement rate: 42.4% | Helped |
| Add IR-AAB edge filtering before ShapeFit | ShapeFit · 12 real-data datasets, excluding Ellis Island and Gendarmenmarkt | mean location error improvement rate: 50.4% | Helped |
| Add 1DSfM edge filtering before ShapeFit | ShapeFit · 12 real-data datasets, excluding Ellis Island and Gendarmenmarkt | mean location error improvement rate: 2.9% | Mixed |
| Use any of the three edge-filtering statistics before a solver | LUD, CLS, or ShapeFit · Ellis Island and Gendarmenmarkt | location error | No clear effect |
| Use IR-AAB rather than naive AAB for corruption classification | Synthetic UC data · Low and moderate noise levels | corruption classification | Helped |
| Use IR-AAB rather than naive AAB or 1DSfM for corruption classification | Synthetic UC data · High noise level, σ=0.2 | corruption classification | No clear effect |
| Reweight AAB inconsistencies iteratively using neighboring-edge suspicion (ablation) | Naive AAB statistic · Synthetic UC(200,0.5,0.2,0) | uncorrupted-edge identification | Helped |
| Sample a small random subset of common neighbors when computing naive AAB (ablation) | Naive AAB · Camera-location estimation | runtime: s=50 | Helped |
Closest papers by meaning
- Cycle-Sync: Robust Global Camera Pose Estimation through Enhanced Cycle-Consistent Synchronization (NeurIPS 2025 · similarity 0.87)
- Tensor-Based Synchronization and the Low-Rankness of the Block Trifocal Tensor (NeurIPS 2024 · similarity 0.84)
- Depth2Pose: A Pose-Based Benchmark for Monocular Depth Estimation without Ground-Truth Depth (NeurIPS 2026 · similarity 0.84)
- Convex Relaxation for Robust Vanishing Point Estimation in Manhattan World (CVPR 2025 · similarity 0.83)
- RUBIK: A Structured Benchmark for Image Matching across Geometric Challenges (CVPR 2025 · similarity 0.83)
- HyperPose: Hypernetwork-Infused Camera Pose Localization and an Extended Cambridge Landmarks Dataset (CVPR 2025 · 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.