Research map / Vision-language and multimodal / Cross-modal learning
Robust Remote Sensing Image-Text Retrieval with Noisy Correspondence
Qiya Song, Yiqiang Xie, Yuan Sun, Renwei Dian, Xudong Kang
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 local contrastive learning (ablation) | RRSITR · RSITMD (80% noise) | mR: 33.38 -> 35.93 | Improved |
| add the multi-modal SPL module (ablation) | RRSITR · RSITMD (80% noise) | mR: 29.80 -> 35.93 | Improved |
| add the robust triplet loss (RTL) (ablation) | RRSITR · RSITMD (80% noise) | mR: 30.21 -> 35.93 | Improved |
| add all three components jointly (ablation) | RRSITR · RSITMD (80% noise) | mR: 32.24 -> 35.93 | Improved |
| use easy-to-hard SPL instead of hard-to-easy (ablation) | RRSITR · RSITMD (80% noise) | mR: 1.22 -> 35.93 | Improved |
| use self-paced weighting instead of random sample weighting (ablation) | RRSITR · RSITMD (80% noise) | mR: 26.52 -> 35.93 | Improved |
| keep ambiguous pairs in SPL instead of removing them (ablation) | RRSITR · RSITMD (80% noise) | mR: 1.10 -> 35.93 | Improved |
| use adaptive soft-margin RTL instead of fixed-margin triplet loss (ablation) | RRSITR · RSITMD (80% noise) | mR: 33.90 -> 35.93 | Improved |
| add local contrastive learning (ablation) | RRSITR · RSITMD (80% noise) | R@1 image-to-text: 15.44 -> 16.90 | Improved |
| replace the fixed-margin triplet loss with RTL (ablation) | RRSITR · RSITMD (80% noise) | text-to-image R@5: 39.45 -> 42.70 | Improved |
| compare RRSITR with CUP as noise rises from 20% to 80% | RRSITR · RSITMD | mR: 45.58 vs 38.99 at 20%; 35.93 vs 27.59 at 80% | Improved |
| compare RRSITR with best baseline on NWPU at 80% noise | RRSITR · NWPU (80% noise) | mR: 12.21 -> 32.13 | Improved |
Closest papers by meaning
- MMM-RS: A Multi-modal, Multi-GSD, Multi-scene Remote Sensing Dataset and Benchmark for Text-to-Image Generation (NeurIPS 2024 · similarity 0.86)
- Remote Sensing Vision-Language Foundation Models without Annotations via Ground Remote Alignment (ICLR 2024 · similarity 0.85)
- SkyScript: A Large and Semantically Diverse Vision-Language Dataset for Remote Sensing (AAAI 2024 · similarity 0.85)
- SkySense-O: Towards Open-World Remote Sensing Interpretation with Vision-Centric Visual-Language Modeling (CVPR 2025 · similarity 0.85)
- Quality-Driven Curation of Remote Sensing Vision-Language Data via Learned Scoring Models (NeurIPS 2025 · similarity 0.85)
- GeoLLaVA-8K: Scaling Remote-Sensing Multimodal Large Language Models to 8K Resolution (NeurIPS 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.
