Research map / 3D vision / Depth and stereo
Bridging the Domain Gap for Ground-to-Aerial Image Matching
Krishna Regmi, 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 joint feature learning with synthesized aerial images | two-stream (Ig, Ia) · CVUSA | top-1% recall: 82.94 -> 92.09 | Improved |
| add feature fusion of synthesized-aerial and ground features | Joint Feature Learning · CVUSA | top-1 recall: 29.75 -> 48.75 | Improved |
| add feature fusion | Joint Feature Learning · CVUSA | top-1% recall: 92.09 -> 95.98 | Improved |
| use synthesized aerial image as query instead of real ground image | two-stream · CVUSA | top-1% recall: 82.94 -> 72.58 | Hurt |
| add multi-scale features (conv_6, conv_7 with conv_8) (ablation) | Feature Fusion · CVUSA | top-1 recall: 8.01 -> 48.75 | Improved |
| do not use global average pooling (ablation) | Feature Fusion · CVUSA | top-1 recall: 16.13 -> 48.75 | Improved |
| do not share weights between ground and aerial encoders (ablation) | Feature Fusion · CVUSA | top-1 recall: 29.94 -> 48.75 | Improved |
| do not share weights between encoders (ablation) | Feature Fusion · CVUSA | top-1% recall: 93.42 -> 95.98 | Improved |
| feature fusion vs CVM-Net-I | Feature Fusion · CVUSA | top-1% recall: 91.4 -> 95.98 | Improved |
| add feature fusion on urban OP dataset | Joint Feature Learning · Orlando-Pittsburgh (OP) | top-1 recall: 38.36 -> 45.57 | Improved |
| add joint feature learning on urban OP dataset | two-stream (Ig, Ia) · Orlando-Pittsburgh (OP) | top-1 recall: 30.61 -> 38.36 | Improved |
| add feature fusion for reverse aerial-to-ground matching | Joint Feature Learning · CVUSA aerial-to-ground | top-1 recall: 27.39 -> 44.99 | Improved |
Closest papers by meaning
- Spatial-Aware Feature Aggregation for Image based Cross-View Geo-Localization (NeurIPS 2019 · similarity 0.90)
- CVM-Net: Cross-View Matching Network for Image-Based Ground-to-Aerial Geo-Localization (CVPR 2018 · similarity 0.88)
- View From Above: Orthogonal-View aware Cross-view Localization (CVPR 2024 · similarity 0.85)
- FG^2: Fine-Grained Cross-View Localization by Fine-Grained Feature Matching (CVPR 2025 · similarity 0.85)
- AerialMegaDepth: Learning Aerial-Ground Reconstruction and View Synthesis (CVPR 2025 · similarity 0.85)
- Multi-Modal Aerial-Ground Cross-View Place Recognition with Neural ODEs (CVPR 2025 · 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.
