Improve model card: add metadata and links to paper/code

README.md

@@ -1,3 +1,30 @@
 ---
 license: cc-by-4.0
+pipeline_tag: image-feature-extraction
 ---
+
+# SDF-Net: Structure-Aware Disentangled Feature Learning for Optical-SAR Ship Re-identification
+
+This repository contains the official weights for **SDF-Net**, presented in the paper [SDF-Net: Structure-Aware Disentangled Feature Learning for Optical-SAR Ship Re-identification](https://huggingface.co/papers/2603.12588).
+
+SDF-Net is a Structure-Aware Disentangled Feature Learning Network that systematically incorporates geometric consistency into optical–SAR ship ReID. Built upon a ViT backbone, it introduces a structure consistency constraint to robustly anchor representations against radiometric variations between passive optical imaging and coherent active radar sensing.
+
+- **Paper:** [https://arxiv.org/abs/2603.12588](https://arxiv.org/abs/2603.12588)
+- **GitHub Repository:** [https://github.com/cfrfree/SDF-Net](https://github.com/cfrfree/SDF-Net)
+
+## Citation
+
+If you find this work useful, please consider citing:
+
+```bibtex
+@article{chen2026sdfnet,
+  title={SDF-Net: Structure-Aware Disentangled Feature Learning for Optical-SAR Ship Re-Identification},
+  author={Chen, Furui and Wang, Han and Sun, Yuhan and You, Jianing and Lv, Yixuan and Zhou, Zhuang and Tan, Hong and Li, Shengyang},
+  journal={arXiv preprint arXiv:2603.12588},
+  year={2026}
+}
+```
+
+## Acknowledgements
+
+This codebase is built upon [TransReID](https://github.com/damo-cv/TransReID) and [TransOSS](https://github.com/Alioth2000/Hoss-ReID). We thank the authors for their excellent work.