Research map / 3D vision / Depth and stereo
A Pose-Only Solution to Visual Reconstruction and Navigation
Qi Cai, Lilian Zhang, Yuanxin Wu, Wenxian Yu, Dewen Hu
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 LiGT linear translation instead of LUD global translation | SfM pipeline · Lund and OpenSLAM (55 data) | running time: approximately 8,000 times on average (140,000 times maximum) | Improved |
| use LiGT instead of LUD for initialization | SfM pipeline · Lund and OpenSLAM (55 data) | memory: approximately 400 times on average (5,000 times maximum) | Improved |
| replace LUD-BA with pose adjustment initialized by LiGT (LiGT-PA) | bundle adjustment · Lund and OpenSLAM (55 data) | running time: approximately 25 times | Improved |
| replace LUD-BA with LiGT-PA | bundle adjustment · Lund and OpenSLAM (55 data) | memory: approximately 15 times | Improved |
| use LiGT instead of LUD initialization | reprojection error before optimization · King's-College | reprojection error: 2493.362 -> 13.880 | Improved |
| use LiGT instead of LUD initialization | reprojection error before optimization · Lund-Cathedral | reprojection error: 3294.693 -> 4.586 | Improved |
| use LiGT as initialization instead of LUD for BA | BA final reprojection error · Basilica-di-SMF | reprojection error: 11.563 -> 0.789 | Improved |
| initialize PA from LiGT instead of LUD | pose adjustment · King's-College | reprojection error: 94.881 -> 0.343 | Improved |
| use LiGT-PA instead of LUD-BA | final optimization · Lund and OpenSLAM | final reprojection error | Improved |
| use LiGT instead of LUD under collinear or small-translation motion | global translation · Lund dataset | 3D scene quality | Improved |
| use LiGT initialization instead of LUD | 3D point scattering before optimization · King's-College | scattering degree (m): 5.292 -> 0.132 | Improved |
| remove point observations of small parallax angle (ablation) | LiGT-BA · Notre Dame (1DSfM) | reconstruction quality | Improved |
Closest papers by meaning
- A Pose-Only Geometric Constraint for Multi-Camera Pose Adjustment (TPAMI 2026 · similarity 0.90)
- GO-SLAM: Global Optimization for Consistent 3D Instant Reconstruction (ICCV 2023 · similarity 0.89)
- Dynamic Visual SLAM using a General 3D Prior (CVPR 2026 · similarity 0.88)
- Dense Dynamic Scene Reconstruction and Camera Pose Estimation from Multi-View Videos (ECCV 2026 · similarity 0.87)
- Photo-SLAM: Real-time Simultaneous Localization and Photorealistic Mapping for Monocular Stereo and RGB-D Cameras (CVPR 2024 · similarity 0.86)
- SLAM3R: Real-Time Dense Scene Reconstruction from Monocular RGB Videos (CVPR 2025 · 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.
