Research map / Label-efficient and robust learning / Self-supervised and contrastive
Geometrized Transformer for Self-Supervised Homography Estimation
Jiazhen Liu, Xirong Li
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 RANSAC-based homography to select updated query points (ablation) | GeoFormer · FIRE | mAUC: 67.3 -> 75.6 | Improved |
| add RANSAC-based homography (ablation) | GeoFormer · HPatches | mAUC: 71.4 -> 76.7 | Improved |
| add focused cross-attention (ablation) | GeoFormer · HPatches | mAUC: 73.0 -> 76.7 | Improved |
| add sparse self-attention (ablation) | GeoFormer · HPatches | mAUC: 75.4 -> 76.7 | Improved |
| use standard instead of linear Transformers (ablation) | GeoFormer · HPatches | mAUC: 73.4 -> 76.7 | Improved |
| use fine-matching2 (symmetric) instead of LoFTR fine-matching (ablation) | GeoFormer · ISC-HE | mAUC: 41.8 -> 44.0 | Improved |
| add GeoFormer to LoFTR | LoFTR · HPatches | AUC@3px vs ASpanFormer: 59.9 -> 68.0 (vs ASpanFormer) | Improved |
| replace LoFTR with GeoFormer for visual localization | LoFTR · Aachen Day-Night (0.25m,2deg) | accuracy: 79.6 -> 77.6 | Hurt |
| add GeoFormer to LoFTR | LoFTR · 640x480 image pair inference | runtime (ms): 77 -> 101 | Hurt |
Closest papers by meaning
- ETO:Efficient Transformer-based Local Feature Matching by Organizing Multiple Homography Hypotheses (NeurIPS 2024 · similarity 0.90)
- Deep Homography Estimation for Visual Place Recognition (AAAI 2024 · similarity 0.88)
- Adapting Dense Matching for Homography Estimation with Grid-based Acceleration (CVPR 2025 · similarity 0.87)
- SSHNet: Unsupervised Cross-modal Homography Estimation via Problem Reformulation and Split Optimization (CVPR 2025 · similarity 0.86)
- HomoFormer: Homogenized Transformer for Image Shadow Removal (CVPR 2024 · similarity 0.86)
- HomoMatcher: Achieving Dense Feature Matching with Semi-Dense Efficiency by Homography Estimation (AAAI 2025 · similarity 0.86)
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.
