Research map / Label-efficient and robust learning / Recognition and long-tail
CODE: Coherence Based Decision Boundaries for Feature Correspondence
Wen-Yan Lin, Fan Wang, Ming‐Ming Cheng, Sai-Kit Yeung, Philip H. S. Torr, Minh N. Do, Jiangbo Lu
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 CODE coherence filtering to A-SIFT | A-SIFT · Own dataset Set A (resized) | F-number: 0.799 -> 0.9027 | Improved |
| add CODE coherence filtering to A-SIFT | A-SIFT · Own dataset Set B (resized, wide baseline) | F-number: 0.6692 -> 0.9552 | Improved |
| add CODE to A-SIFT: number of matches | A-SIFT · Own dataset Set A (resized) | average match number: 396.84 -> 7.3447e+03 | Improved |
| add CODE to A-SIFT: failures | A-SIFT · Own dataset Set B (resized) | total failures: 3/7 -> 0/7 | Improved |
| add CODE to A-SIFT: coverage area | A-SIFT · Own dataset Set B (resized) | Area %: 20.9% -> 100% | Improved |
| use CODE instead of Mode Seeking | CODE vs Mode Seeking · Co-recognition dataset | F-number (area): 0.5901 -> 0.7249 | Improved |
| use CODE instead of NRDC(t) | CODE vs NRDC(t) · Co-recognition dataset | F-number (area): 0.6155 -> 0.7249 | Improved |
| use CODE instead of NRDC | CODE vs NRDC · Own dataset Set A (full resolution) | F-number: 0.803 -> 0.940 | Improved |
| use CODE instead of NRDC | CODE vs NRDC · Own dataset Set B (full resolution) | F-number: 0.822 -> 0.960 | Improved |
| use CODE instead of Mode Seeking | CODE vs Mode Seeking · ETHZ-toys | F-number: 0.3767 -> 0.9762 | Improved |
| use CODE instead of NRDC | CODE vs NRDC · ETHZ-toys | recall: 48.8% -> 95.4% | Improved |
| use CODE rather than NRDC for dense matching | CODE vs NRDC · Own dataset Set A (full resolution) | Av. Match Num: 4.8556e+04 -> 4.41e+03 | Hurt |
Closest papers by meaning
- MATCHA: Towards Matching Anything (CVPR 2025 · similarity 0.83)
- Noisy Correspondence Rectification via Asymmetric Similarity Learning (AAAI 2025 · similarity 0.83)
- Mitigating Noisy Correspondence by Geometrical Structure Consistency Learning (CVPR 2024 · similarity 0.82)
- Diffusion Model for Dense Matching (ICLR 2024 · similarity 0.82)
- Disentangled Representation Learning with the Gromov-Monge Gap (ICLR 2025 · similarity 0.82)
- ReCon: Enhancing True Correspondence Discrimination through Relation Consistency for Robust Noisy Correspondence Learning (CVPR 2025 · 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.
