Research map / Vision-language and multimodal / Multimodal LLMs
RSICCLLM: A Multimodal Large Language Model for Remote Sensing Image Change Captioning
Yelin Wang, Zijia Song, Shuo Ye, Chuanguang Yang, Miaoyu Wang, Yong Xu, Zhulin An, Yongjun Xu, Zitong Yu
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 difference map T (ablation) | RSICCLLM-sft · RSICI in-domain | BLEU-4: 60.45 -> 62.78 | Improved |
| add delta S convolutional change cues (ablation) | RSICCLLM · RSICI in-domain | BLEU-4: 61.90 -> 62.78 | Improved |
| add geometric-texture statistics G (ablation) | RSICCLLM · RSICI in-domain | BLEU-4: 62.54 -> 62.78 | Improved |
| add DNPO after SFT (ablation) | RSICCLLM · RSICI in-domain | BLEU-4: 61.68 -> 62.78 | Improved |
| add SFT to base Qwen2.5-VL-7B (ablation) | Qwen2.5-VL-7B · RSICI in-domain | BLEU-4: 3.21 -> 61.68 | Improved |
| combine info-filtered and replacement negatives (ablation) | DNPO · RSICI in-domain | BLEU-4: 62.38 -> 62.78 | Improved |
| use cross-attention instead of add fusion for delta S (ablation) | RSICCLLM-sft · RSICI in-domain | BLEU-1: 79.14 -> 81.28 | Improved |
| add KL stabilizer L_S to DPO (ablation) | DNPO · RSICI in-domain | BLEU-4 and ROUGE-1 | Improved |
| compare 7B RSICCLLM to larger general VLMs | RSICCLLM · RSICI out-of-domain | SBS: 71.18 -> 73.86 | Improved |
| use Difference-aware SFT instead of standard SFT at small data (ablation) | RSICCLLM · RSICI in-domain | BLEU-4 / ROUGE-1 | 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.
- TEOChat-7B — TEOChat: A Large Vision-Language Assistant for Temporal Earth Observation Data (ICLR 2025) · better on 3 of 3 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
- RSRCC: A Remote Sensing Regional Change Comprehension Benchmark Constructed via Retrieval-Augmented Best-of-𝑁 Ranking (NeurIPS 2026 · similarity 0.92)
- RSCC: A Large-Scale Remote Sensing Change Caption Dataset for Disaster Events (NeurIPS 2025 · similarity 0.89)
- Quality-Driven Curation of Remote Sensing Vision-Language Data via Learned Scoring Models (NeurIPS 2025 · similarity 0.87)
- VRSBench: A Versatile Vision-Language Benchmark Dataset for Remote Sensing Image Understanding (NeurIPS 2024 · similarity 0.87)
- MMM-RS: A Multi-modal, Multi-GSD, Multi-scene Remote Sensing Dataset and Benchmark for Text-to-Image Generation (NeurIPS 2024 · similarity 0.87)
- GeoChat: Grounded Large Vision-Language Model for Remote Sensing (CVPR 2024 · 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.
