Research map / Detection and segmentation / CNN features
Dual-Resolution Correspondence Networks
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 coarse-resolution consensus scores to guide fine-resolution matching | DualRC-Net · HPatches | MMA | Helped |
| Use coarse-resolution consensus scores to guide fine-resolution matching | DualRC-Net · InLoc | relocalization performance | Helped |
| Use coarse-resolution consensus scores to guide fine-resolution matching | DualRC-Net · Aachen Day-Night | pose estimation performance | Helped |
| Use FPN variant 1, directly fusing conv4_23 with conv2_3 (ablation) | DualRC-Net · HPatches illumination-change sequences | MMA | Hurt |
| Use FPN variant 1, directly fusing conv4_23 with conv2_3 (ablation) | DualRC-Net · InLoc | relocalization performance | Hurt |
| Use FPN variant 2 with an additional 3×3 convolution module (ablation) | DualRC-Net · HPatches and InLoc | matching performance | No clear effect |
| Use FPN variant 3 with its alternative additional 3×3 convolution module (ablation) | DualRC-Net · HPatches | MMA | Helped |
| Use FPN variant 3 with its alternative additional 3×3 convolution module (ablation) | DualRC-Net · InLoc | relocalization performance | No clear effect |
| Add conv5_3 to the FPN feature fusion (variant 4) (ablation) | DualRC-Net · HPatches | MMA | Hurt |
| Increase output channels from 256 to 1024 in the default FPN (ablation) | DualRC-Net · HPatches illumination-change sequences | MMA | Helped |
| Increase output channels from 256 to 1024 in FPN variant 4 (ablation) | DualRC-Net · HPatches | MMA | No clear effect |
Closest papers by meaning
- Correspondence Networks With Adaptive Neighbourhood Consensus (CVPR 2020 · similarity 0.88)
- GLU-Net: Global-Local Universal Network for Dense Flow and Correspondences (CVPR 2020 · similarity 0.87)
- Correspondence Transformers With Asymmetric Feature Learning and Matching Flow Super-Resolution (CVPR 2023 · similarity 0.86)
- GOCor: Bringing Globally Optimized Correspondence Volumes into Your Neural Network (NeurIPS 2020 · similarity 0.86)
- FC-GNN: Recovering Reliable and Accurate Correspondences from Interferences (CVPR 2024 · similarity 0.85)
- Efficient LoFTR: Semi-Dense Local Feature Matching with Sparse-Like Speed (CVPR 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.