Research map / Learning theory and algorithms / Matrix and tensor methods
Registration beyond Points: General Affine Subspace Alignment via Geodesic Distance on Grassmann Manifold
Jaeho Shin, Hyeonjae Gil, Junwoo Jang, Maani Ghaffari, Ayoung Kim
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 |
|---|---|---|---|
| refine LinEq solution with Grassmann geodesic cost | LinEq (Plucker linear solver) · Replica Office3 RGB-D odometry | APE translation (m): 0.40 -> 0.09 | Improved |
| refine LinEq solution with Grassmann geodesic cost | LinEq (Plucker linear solver) · Replica Office3 RGB-D odometry | APE rotation (deg): 7.06 -> 3.25 | Improved |
| refine LinEq solution with Grassmann geodesic cost | LinEq · Replica Room1 RGB-D odometry | APE translation (m): 0.91 -> failure | Hurt |
| use geodesic cost refinement of ASP3L on chessboard | ASP3L · 6D chessboard pose | translation error (mm): 1.91 -> 1.13 | Improved |
| refine ASP3L with our cost for rotation | ASP3L · 6D chessboard pose | rotation error (deg): 0.30 -> 0.38 | Mixed |
| replace point-to-plane linear translation fit with Grassmann cost under BnB rotation | RoPnL · 6D chessboard pose | rotation error (deg): 0.36 -> 0.25 | Improved |
| use geodesic cost BnB for translation instead of linear fit | RoPnL · 6D chessboard pose | translation error (mm): 1.43 -> 1.23 | Mixed |
| refine LinEq plane registration with our cost | LinEq · Space Station object registration | translation error | Improved |
| add Refine/BnB on top of Olsson point-based method | Olsson · Space Station object registration | rotation error | Improved |
| use Grassmann cost with BnB vs Park colored point cloud method | Park · Replica Office3 RGB-D odometry | APE translation (m): 1.37 -> 0.11 | Improved |
| Refine trades rotational accuracy for translation accuracy vs BnB | BnB · Space Station object registration | rotation/translation error | Mixed |
| add correspondence search to BnB solver | BnB · Correspondence-free localization | pose error: 0.65 deg and 0.03% error | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Park — ParK: Sound and Efficient Kernel Ridge Regression by Feature Space Partitions (NeurIPS 2021) · better on 5 of 5 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Approximate Grassmannian Intersections: Subspace-Valued Subspace Learning (ICCV 2017 · similarity 0.88)
- Jumping Manifolds: Geometry Aware Dense Non-Rigid Structure From Motion (CVPR 2019 · similarity 0.87)
- Making Affine Correspondences Work in Camera Geometry Computation (ECCV 2020 · similarity 0.87)
- Non-Rigid Point Set Registration with Robust Transformation Estimation under Manifold Regularization (AAAI 2017 · similarity 0.86)
- Scalable 3D Registration via Truncated Entry-wise Absolute Residuals (CVPR 2024 · similarity 0.84)
- Sample-Efficient Geometry Reconstruction from Euclidean Distances using Non-Convex Optimization (NeurIPS 2024 · similarity 0.84)
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.
