Research map / Learning theory and algorithms / Optimal transport
Robust Point Matching with Distance Profiles
YoonHaeng Hur, Yuehaw Khoo
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 |
|---|---|---|---|
| add thresholding to keep only the best 50% of correspondences (ablation) | distance profile matching alignment · Cryo-EM human gamma-secretase | Wasserstein-2 distance (lower is better): 10.68 -> 9.04 | Improved |
| tighten threshold from best 50% to best 10% of correspondences (ablation) | distance profile matching alignment · Cryo-EM human gamma-secretase | Wasserstein-2 distance (lower is better): 9.04 -> 8.31 | Improved |
| add thresholding to keep only the best 50% of correspondences (ablation) | distance profile matching alignment · Cryo-EM voltage-gated calcium channel Cav1.1 | Wasserstein-2 distance (lower is better): 11.10 -> 8.23 | Improved |
| tighten threshold from best 50% to best 10% of correspondences (ablation) | distance profile matching alignment · Cryo-EM voltage-gated calcium channel Cav1.1 | Wasserstein-2 distance (lower is better): 8.23 -> 9.17 | Hurt |
| add one outlier component (t=s=K+1) (ablation) | distance profile matching · Gaussian mixture, d=3, K=10 | probability of perfect matching | Hurt |
| increase sample size from n=m=500 to n=m=1000 (ablation) | distance profile matching · Gaussian mixture, d=3, K=10 | probability of perfect matching | Improved |
| increase dimension from d=3 to d=10 (ablation) | distance profile matching · Gaussian mixture, K=10, t=s=11 | matching accuracy | Hurt |
| use distance profile matching instead of standard OT under a rotation changing all coordinates | point matching under rigid transformation · Noise stability, d=10, n=100 | probability of recovering true permutation | Improved |
| use distance profile matching instead of standard OT under a rotation affecting only two coordinates | point matching under rigid transformation · Noise stability, d=10, n=100 | probability of recovering true permutation | No clear effect |
| raise proportion of outlier components (K=20, t=s=24 vs t=s=22) (ablation) | distance profile matching · Gaussian mixture, d=3 | matching accuracy | Hurt |
Closest papers by meaning
- Best-Buddies Similarity—Robust Template Matching Using Mutual Nearest Neighbors (TPAMI 2017 · similarity 0.87)
- Probabilistic Model for Robust Affine and Non-Rigid Point Set Matching (TPAMI 2016 · similarity 0.87)
- Locality Preserving Refinement for Shape Matching with Functional Maps (AAAI 2024 · similarity 0.86)
- Sample-Efficient Geometry Reconstruction from Euclidean Distances using Non-Convex Optimization (NeurIPS 2024 · similarity 0.85)
- Partial Gromov-Wasserstein Metric (ICLR 2025 · similarity 0.85)
- Wormhole Loss for Partial Shape Matching (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.
