Research map / Video / Video segmentation and tracking
GAMa: Cross-view Video Geo-localization
Shruti Vyas, Chen Chen, 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 transformer encoder on aerial images (ablation) | CBn with video transformer · GAMa test (day videos, CN aerial) | R@1: 14.1 -> 15.2 | Improved |
| add transformer encoder on ground video (ablation) | CBn-CN · GAMa test (day videos, CN aerial) | R@1: 11.5 -> 14.1 | Improved |
| add transformer on both video and aerial instead of aerial only (ablation) | GAMa-Net · GAMa test (day videos, CN aerial) | R@1: 15.2 -> 14.6 | Hurt |
| add hierarchical gallery reduction to top 1% (ablation) | GAMa-Net · GAMa test (day videos, CN aerial) | R@1: 15.2 -> 18.3 | Improved |
| add hierarchical gallery reduction to top 1% (ablation) | GAMa-Net · GAMa test (day videos, CN aerial) | Top-1@1.0 mile: 36.1 -> 43.6 | Improved |
| reduce gallery to top 10 instead of top 1% (ablation) | GAMa-Net hierarchical · GAMa test (day videos, CN aerial) | R@1: 18.3 -> 16.2 | Hurt |
| use contrastive loss instead of triplet loss (ablation) | CBn vs Triplet-Bn · GAMa test (day videos, CN aerial) | R@1: <1 -> 11.5 | Improved |
| use 2D-CNN instead of 3D-CNN for aerial sequence of 8 (ablation) | Screening network · GAMa video-level screening network | R@1: 7.5 -> 12.2 | Improved |
| lengthen aerial sequence from 8 to 32 (ablation) | Screening network 2D-CNN · GAMa video-level screening network | R@1: 12.2 -> 8.2 | Hurt |
| train on full day and night data (ablation) | Hierarchical · GAMa test (day videos, CN aerial) | R@1: 18.3 -> 19.4 | Improved |
| add augmentations (combined model) to CBn-CN (ablation) | CBn-CN · GAMa test (day videos, CN aerial) | R@1: 11.5 -> 11.6 | No clear effect |
| use video input instead of single frame | Image-CBn vs CBn · GAMa test (day videos, CN aerial) | R@1: 9.5 -> 11.5 | Improved |
Closest papers by meaning
- GAReT: Cross-view Video Geolocalization with Adapters and Auto-Regressive Transformers (ECCV 2024 · similarity 0.89)
- Lending Orientation to Neural Networks for Cross-View Geo-Localization (CVPR 2019 · similarity 0.88)
- Where Am I Looking At? Joint Location and Orientation Estimation by Cross-View Matching (CVPR 2020 · similarity 0.88)
- Aligning Geometric Spatial Layout in Cross-View Geo-Localization via Feature Recombination (AAAI 2024 · similarity 0.87)
- Cross-View Geo-Localization via Learning Disentangled Geometric Layout Correspondence (AAAI 2023 · similarity 0.87)
- From Coarse to Fine: A Matching and Alignment Framework for Unsupervised Cross-View Geo-Localization (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.
