Research map / Vision-language and multimodal / Cross-modal learning
MM-OVSeg: Multimodal Optical-SAR Fusion for Open-Vocabulary Segmentation in Remote Sensing
Yimin Wei, Aoran Xiao, Hongruixuan Chen, Junshi Xia, Naoto Yokoya
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 Cross-Modal Unification (SAR-RGB DINO alignment) to the full model (ablation) | MM-OVSeg · DDHR-SK (cloudy) | mIoU: 64.1 -> 73.1 | Improved |
| add Dual-Encoder Fusion on top of the optical-only baseline (ablation) | MM-OVSeg · DDHR-SK (cloudy) | mIoU: 55.0 -> 64.1 | Improved |
| use InfoNCE rather than MSE for the CMU alignment loss (ablation) | MM-OVSeg CMU · DDHR-SK (cloudy) | mIoU: 67.7 -> 73.1 | Improved |
| use InfoNCE rather than L1 for the CMU alignment loss (ablation) | MM-OVSeg CMU · DDHR-SK (cloudy) | mIoU: 69.0 -> 73.1 | Improved |
| any CMU alignment loss (MSE, L1) versus no CMU (ablation) | MM-OVSeg · DDHR-SK (cloudy) | mIoU: 64.1 -> 67.7 (MSE) | Improved |
| add SAR input via multimodal fusion to an optical-only open-vocabulary model | MM-OVSeg vs GSNet · PIE-clean (clear sky) | mIoU: 57.2 -> 59.7 | Improved |
| multimodal Optical-SAR fusion versus best unimodal method | MM-OVSeg vs GSNet · Six benchmarks (mean) | mIoU: 45.6 -> 51.7 | Improved |
| multimodal fusion for cross-domain transfer | MM-OVSeg vs CAT-Seg · DDHR-SK to DDHR-CH | mIoU: 27.8 -> 42.6 | Improved |
| fuse SAR for the unseen water class (ablation) | MM-OVSeg · DDHR-SK (cloudy) | IoU (Water): 19.1 -> 84.8 | Improved |
| add CMU alone without DEF effect on Forest class (removing CMU) (ablation) | MM-OVSeg · DDHR-SK (cloudy) | IoU (Forest): 57.2 -> 87.3 | 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.
- CAT-Seg — CAT-Seg: Cost Aggregation for Open-Vocabulary Semantic Segmentation (CVPR 2024) · 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
- Towards Open-Vocabulary Remote Sensing Image Semantic Segmentation (AAAI 2025 · similarity 0.86)
- FUSAR-GPT: A Spatiotemporal Feature-Embedded and Two-Stage Decoupled Visual Language Model for SAR Imagery (CVPR 2026 · similarity 0.86)
- SkySense-O: Towards Open-World Remote Sensing Interpretation with Vision-Centric Visual-Language Modeling (CVPR 2025 · similarity 0.84)
- Bridging Remote Sensors with Multisensor Geospatial Foundation Models (CVPR 2024 · similarity 0.84)
- M3LEO: A Multi-Modal, Multi-Label Earth Observation Dataset Integrating Interferometric SAR and Multispectral Data (NeurIPS 2024 · similarity 0.84)
- SkySense: A Multi-Modal Remote Sensing Foundation Model Towards Universal Interpretation for Earth Observation Imagery (CVPR 2024 · 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.
