Research map / Image generation / Autoregressive and latent generation
Using Spatial Order to Boost the Elimination of Incorrect Feature Matches
Check 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 |
|---|---|---|---|
| Use the K1D overlap-aware estimate instead of estimating NG on the full sequence (ablation) | Correct-match count estimation · Syn1 and Syn2 | µ(NG) (%): 14.5 -> 4; 10.4 -> 3.7 | Helped |
| Use K1D rather than joint interval sampling with K2D (ablation) | Correct-match count estimation · Syn1 and Syn2 | Runtime (ms): 83.2 -> 11.2; 79.2 -> 11.1 | Helped |
| Estimate NG on the full sequence without accounting for non-overlapping margins (ablation) | Correct-match count estimation · MPar14 and MPar05&06 | µ(NG) (%): 11.5 -> 7.1; 23.1 -> 6.7 | Hurt |
| Add the estimated-NG halting condition to USAC | USAC RANSAC · All evaluation datasets | Runtime: total runtime with the halting condition is 69% of the total runtime without the halting condition | Helped |
| Discard image pairs below an estimated NG threshold before running RANSAC | SfM image matching on Barcelona · Barcelona | Runtime: 7%; 1.2% | Helped |
| Discard image pairs below an estimated NG threshold before running RANSAC | SfM image matching on Barcelona · Barcelona | Correct inliers retained (%): 87%; 80% | Hurt |
| Rank matches using the combined spatial-order and Lowe probabilities rather than either probability alone | Guided RANSAC match sampling · All datasets except Barcelona | Inlier precision | Mixed |
| Estimate NG while accounting for partial-overlap margins (ablation) | Correct-match count estimation · Middlebury partial-overlap datasets | µ(NG) (%): 11.5 -> 7.1; 23.1 -> 6.7 | Helped |
Closest papers by meaning
- MATCHA: Towards Matching Anything (CVPR 2025 · similarity 0.84)
- HomoMatcher: Achieving Dense Feature Matching with Semi-Dense Efficiency by Homography Estimation (AAAI 2025 · similarity 0.84)
- OmniGlue: Generalizable Feature Matching with Foundation Model Guidance (CVPR 2024 · similarity 0.84)
- RUBIK: A Structured Benchmark for Image Matching across Geometric Challenges (CVPR 2025 · similarity 0.83)
- SegMASt3R: Geometry Grounded Segment Matching (NeurIPS 2025 · similarity 0.83)
- Partial-to-Partial Shape Matching with Geometric Consistency (CVPR 2024 · similarity 0.82)
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.