Research map / Image generation / Autoregressive and latent generation
Large-Scale Image Geo-Localization Using Dominant Sets
Eyasu Zemene, Yonatan Tariku Tesfaye, Haroon Idrees, Andrea Prati, Marcello Pelillo, Mubarak Shah
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 constrained dominant set post-processing (ablation) | NetVLAD · 102k Google street view images | rank-1 accuracy (%): 49.2 -> 51.60 | Improved |
| add constrained dominant set post-processing (ablation) | NetVLAD* · 102k Google street view images | rank-1 accuracy (%): 56.00 -> 58.05 | Improved |
| add constrained dominant set post-processing (ablation) | RMAC · 102k Google street view images | rank-1 accuracy (%): 35.16 -> 40.18 | Improved |
| add constrained dominant set post-processing (ablation) | MAC · 102k Google street view images | rank-1 accuracy (%): 29.00 -> 36.30 | Improved |
| add constrained dominant set post-processing (ablation) | NetVLAD · WorldCities | rank-1 accuracy (%): 50.00 -> 53.04 | Improved |
| add constrained dominant set post-processing (ablation) | NetVLAD* · WorldCities | rank-1 accuracy (%): 50.61 -> 52.23 | Improved |
| add constrained dominant set post-processing (ablation) | RMAC · WorldCities | rank-1 accuracy (%): 29.96 -> 33.16 | Improved |
| add constrained dominant set post-processing (ablation) | MAC · WorldCities | rank-1 accuracy (%): 22.87 -> 26.31 | Improved |
| add constrained dominant set post-processing to the voting baseline (ablation) | dominant set geo-localization · 102k Google street view images | localization within error threshold: about 10% -> about 20% improvement over state of the art | Improved |
| add constrained dominant set post-processing to the voting baseline (ablation) | dominant set geo-localization · WorldCities | localization within error threshold: around 2% -> 7% improvement | Improved |
| use dominant-set feature matching instead of GMCP as the number of nearest neighbors grows (ablation) | feature matching · 102k Google street view images | accuracy within 30 m | Improved |
| fuse all global features (fine-tuned NetVLAD, CNN7, CNN6, GIST, HSV) with dynamic weights instead of single features (ablation) | post-processing · test set | localization accuracy | Improved |
Closest papers by meaning
- Where We Are and What We're Looking At: Query Based Worldwide Image Geo-Localization Using Hierarchies and Scenes (CVPR 2023 · similarity 0.89)
- Cross-View Image Matching for Geo-Localization in Urban Environments (CVPR 2017 · similarity 0.89)
- VIGOR: Cross-View Image Geo-Localization Beyond One-to-One Retrieval (CVPR 2021 · similarity 0.88)
- GeoRanker: Distance-Aware Ranking for Worldwide Image Geolocalization (NeurIPS 2025 · similarity 0.88)
- From Coarse to Fine: A Matching and Alignment Framework for Unsupervised Cross-View Geo-Localization (AAAI 2025 · similarity 0.87)
- Aligning Geometric Spatial Layout in Cross-View Geo-Localization via Feature Recombination (AAAI 2024 · similarity 0.85)
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.
