Research map / Vision-language and multimodal / Human and embodied modeling
Where am I? Cross-View Geo-localization with Natural Language Descriptions
Junyan Ye, Honglin Lin, Leyan Ou, Dairong Chen, Zihao Wang, Qi Zhu, Conghui He, Weijia 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 Expanded Positional Embedding (EPE) (ablation) | CLIP, satellite · CVG-Text satellite New York | R@1 (%): 35.08 -> 46.25 | Improved |
| add Expanded Positional Embedding (EPE) (ablation) | CLIP, OSM · CVG-Text OSM New York | R@1 (%): 31.50 -> 59.08 | Improved |
| add Expanded Positional Embedding (EPE) (ablation) | SigLIP, satellite · CVG-Text satellite New York | R@1 (%): 19.67 -> 29.50 | Improved |
| add Expanded Positional Embedding (EPE) (ablation) | SigLIP, OSM · CVG-Text OSM New York | R@1 (%): 20.17 -> 45.25 | Improved |
| add Explainable Retrieval Module confidence re-ranking (ablation) | CrossText2Loc · CVG-Text satellite New York | R@1 (%): 46.25 -> 50.33 | Improved |
| add Explainable Retrieval Module confidence re-ranking (ablation) | CrossText2Loc · CVG-Text OSM Tokyo | R@1 (%): 34.33 -> 36.92 | Improved |
| add Explainable Retrieval Module confidence re-ranking (ablation) | CrossText2Loc · CVG-Text OSM Brisbane | R@1 (%): 46.08 -> 48.75 | Improved |
| use OCR and open-world segmentation assisted text instead of direct GPT-4o text (ablation) | CrossText2Loc · CVG-Text OSM New York | R@1 (%): 25.17 -> 59.08 | Improved |
| use OCR and open-world segmentation assisted text instead of direct GPT-4o text (ablation) | CrossText2Loc · CVG-Text satellite New York | R@1 (%): 38.00 -> 46.25 | Improved |
| add text query branch to image-based retrieval | Sample4G · CVG-Text New York OSM | R@1 (%): 27.10 -> 67.30 | Improved |
| add text query branch to image-based retrieval | Sample4G · CVG-Text New York satellite | R@1 (%): 91.70 -> 98.40 | Improved |
| train CrossText2Loc instead of baseline CLIP | CLIP-L/14 · CVG-Text satellite New York | R@1 (%): 35.08 -> 46.25 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- BLIP — BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation (ICML 2022) · better on 6 of 6 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- VIGOR: Cross-View Image Geo-Localization Beyond One-to-One Retrieval (CVPR 2021 · similarity 0.89)
- From Coarse to Fine: A Matching and Alignment Framework for Unsupervised Cross-View Geo-Localization (AAAI 2025 · similarity 0.87)
- Semantic Ambiguity Modeling and Propagation for Fine-Grained Visual Cross View Geo-Localization (AAAI 2025 · similarity 0.86)
- GeoReasoner: Geo-localization with Reasoning in Street Views using a Large Vision-Language Model (ICML 2024 · similarity 0.85)
- Recognition through Reasoning: Reinforcing Image Geo-localization with Large Vision-Language Models (NeurIPS 2025 · similarity 0.85)
- Unleashing Unlabeled Data: A Paradigm for Cross-View Geo-Localization (CVPR 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.
