Research map / Graph learning / Graph algorithms and clustering
Searching for Representative Modes on Hypergraphs for Robust Geometric Model Fitting
Hanzi Wang, Guobao Xiao, Yan Yan, David Suter
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 the neighboring constraint (MSHF2 vs MSHF1) to reduce computation (ablation) | MSHF · Synthetic circle fitting (16 circles) | CPU time (s): 9.46 -> 1.47 | Improved |
| add the neighboring constraint (MSHF2 vs MSHF1) on fitting accuracy (ablation) | MSHF · AdelaideRMF homography segmentation | mean fitting error (%): 7.38 -> 7.38 | No clear effect |
| add the neighboring constraint (MSHF2 vs MSHF1) to reduce computation (ablation) | MSHF · AdelaideRMF motion segmentation | CPU time | Improved |
| add hypergraph reduction before mode-seeking (G* instead of G) (ablation) | MSHF · Homography segmentation on Neem | number of model instances found | Improved |
| replace MSH with MSHF (keep more vertices) | MSHF1/MSHF2 · AdelaideRMF homography segmentation | mean fitting error (%): 8.70 -> 7.38 | Improved |
| replace MSH with MSHF | MSHF1/MSHF2 · AdelaideRMF motion segmentation | mean fitting error (%): 8.17 -> 7.41 | Improved |
| use MSHF instead of MSH on 3 lines synthetic data | MSHF1/MSHF2 · Synthetic line fitting (3 lines) | avg fitting error (%): 0.16 -> 0.14 | Improved |
| use MSHF instead of MSH on 16 circles synthetic data | MSHF1/MSHF2 · Synthetic circle fitting (16 circles) | avg fitting error (%): 4.46 -> 3.59 | Improved |
| use MSHF instead of T-linkage on Tracks line fitting for speed | MSHF2 · Real image line fitting (Tracks) | CPU time (s): 23,256.00 -> 7.12 | Improved |
| use MSHF instead of MSH when the mode search drops low-support modes | MSHF · Synthetic circle fitting (3 circles) | std of fitting error: 6.33 -> 0.63 | Improved |
| use MSHF where a plane has few model hypotheses (4.17% inliers) | MSHF2 · Homography segmentation on Unionhouse | number of model instances found | Hurt |
| increase number of generated model hypotheses | MSH/MSHF · AdelaideRMF | fitting error and model count | Improved |
Closest papers by meaning
- Hypergraph Optimization for Multi-Structural Geometric Model Fitting (AAAI 2019 · similarity 0.92)
- MultiLink: Multi-Class Structure Recovery via Agglomerative Clustering and Model Selection (CVPR 2021 · similarity 0.86)
- Geometric Multi-Model Fitting With a Convex Relaxation Algorithm (CVPR 2018 · similarity 0.86)
- Higher-Order Multicuts for Geometric Model Fitting and Motion Segmentation (TPAMI 2022 · similarity 0.85)
- A CUR Decomposition-Based Mix-Order Framework for Large-Scale Hypergraph Matching (TPAMI 2026 · similarity 0.85)
- Hypergraph Learning for Unsupervised Graph Alignment via Optimal Transport (AAAI 2025 · 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.
