Research map / Detection and segmentation / Open-set and few-shot detection
Breaking Rectangular Shackles: Cross-View Object Segmentation for Fine-Grained Object Geo-Localization
Qingwang Zhang, Yingying Zhu
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 SAM Prompt Stage (ablation) | TROGeo · CVOGL-Seg Drone->Satellite (val) | mIoU: 39.62 -> 55.54 | Improved |
| add SAM Prompt Stage | DetGeo · CVOGL-Seg Drone->Satellite (test) | mIoU: 30.50 -> 42.27 | Improved |
| add SAM Prompt Stage | Sample4Geo · CVOGL-Seg Ground->Satellite (test) | mIoU: 5.92 -> 6.26 | Mixed |
| add CVOPM (ablation) | TROGeo · CVOGL-Seg Drone->Satellite (val) | mIoU: 35.45 -> 55.54 | Improved |
| add CVOPM (ablation) | TROGeo · CVOGL Drone->Satellite (val) | Acc@50%: 43.88% -> 66.63% | Improved |
| add object segmentation task OST (ablation) | TROGeo · CVOGL-Seg Drone->Satellite (val) | mIoU: 54.09 -> 55.54 | Improved |
| add object segmentation task OST (ablation) | TROGeo · CVOGL Drone->Satellite (val) | Acc@50%: 65.87% -> 66.63% | Improved |
| share weights between query and reference encoders (ablation) | TROGeo · CVOGL-Seg Drone->Satellite (val) | mIoU: 37.91 -> 55.54 | Improved |
| share weights between query and reference encoders (ablation) | TROGeo · CVOGL Ground->Satellite (val) | Acc@50%: 44.53% -> 46.59% | Improved |
| set alpha to 2 instead of 4 (ablation) | TROGeo · CVOGL-Seg Drone->Satellite (val) | mIoU: 53.37 -> 55.54 | Improved |
| use segmentation masks instead of detection boxes as output | TROGeo vs DetGeo · CVOGL-Seg Drone->Satellite (test) | mIoU: 30.50 -> 56.59 | Improved |
| use bbox+point prompts for SAM instead of mask prompts (ablation) | SPS · CVOGL-Seg | mIoU | Improved |
Closest papers by meaning
- Recurrent Cross-View Object Geo-Localization (ECCV 2026 · similarity 0.91)
- OpenCVL: An Open, Diverse, and Large-Scale Dataset for Fine-Grained Cross-View Localization (ECCV 2026 · similarity 0.89)
- MOGeo: Beyond One-to-One Cross-View Object Geo-localization (CVPR 2026 · similarity 0.89)
- Aligning Geometric Spatial Layout in Cross-View Geo-Localization via Feature Recombination (AAAI 2024 · similarity 0.87)
- Unleashing Unlabeled Data: A Paradigm for Cross-View Geo-Localization (CVPR 2024 · similarity 0.87)
- From Coarse to Fine: A Matching and Alignment Framework for Unsupervised Cross-View Geo-Localization (AAAI 2025 · similarity 0.87)
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.
