Research map / Graph learning / Graph algorithms and clustering
Progressive Neighbor Consistency Mining for Correspondence Pruning
Read the paperCheck 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 NCMNet instead of the strongest baseline | correspondence pruning · YFCC100M known | mAP5° (%): 41.33 -> 52.33 | Helped |
| use NCMNet instead of the strongest baseline | correspondence pruning · YFCC100M unknown | mAP5° (%): 53.10 -> 63.43 | Helped |
| use NCMNet instead of the strongest baseline | correspondence pruning · SUN3D known | mAP5° (%): 20.86 -> 26.12 | Helped |
| use NCMNet instead of the strongest baseline | correspondence pruning · SUN3D unknown | mAP5° (%): 17.84 -> 20.66 | Helped |
| add the Self-Context Extraction layer (ablation) | iterative-pruning correspondence network · YFCC100M unknown | mAP5°: 53.10 -> 56.50 | Helped |
| add the Cross-Context Interaction layer (ablation) | NCMNet with SCE · YFCC100M unknown | mAP5°: 56.50 -> 58.63 | Helped |
| add progressive neighbor refinement (ablation) | NCMNet with SCE and CCI · YFCC100M unknown | mAP5°: 58.63 -> 61.73 | Helped |
| add the Order-Aware block (ablation) | NCMNet with progressive neighbor refinement · YFCC100M unknown | mAP5°: 61.73 -> 63.43 | Helped |
| combine spatial, feature-space, and global-graph neighbors rather than using one type (ablation) | Neighbor Consistency block · YFCC100M unknown | mAP5°: 61.40; 62.60; 61.73 -> 63.43 | Helped |
| replace grouped convolution with average-pooling and MLPs (ablation) | SCE layer · YFCC100M unknown | mAP5°: 63.43 -> 61.48 | Hurt |
| replace grouped convolution with max-pooling and MLPs (ablation) | SCE layer · YFCC100M unknown | mAP5°: 63.43 -> 62.75 | Hurt |
| replace grouped convolution with a 1 × k convolution (ablation) | SCE layer · YFCC100M unknown | mAP5°: 63.43 -> 62.88 | Hurt |
Closest papers by meaning
- Graph Context Transformation Learning for Progressive Correspondence Pruning (AAAI 2024 · similarity 0.87)
- VSFormer: Visual-Spatial Fusion Transformer for Correspondence Pruning (AAAI 2024 · similarity 0.86)
- BCLNet: Bilateral Consensus Learning for Two-View Correspondence Pruning (AAAI 2024 · similarity 0.85)
- FC-GNN: Recovering Reliable and Accurate Correspondences from Interferences (CVPR 2024 · similarity 0.84)
- FastJAM: a Fast Joint Alignment Model for Images (NeurIPS 2025 · similarity 0.84)
- Locality Preserving Refinement for Shape Matching with Functional Maps (AAAI 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.