GSV2X: Geometry-Aware Uncertainty Modeling and Orthogonal Fusion for Robust Roadside Perception

Accepted to CVPR 2026

Paper • Code • Citation

---

✨ Overview

GSV2X is a research project accepted by CVPR 2026.

TODO: Add a 2–4 sentence summary of the problem, key contributions, and main results.

---

🗞️ News

• 2026-03-27: Project page initialized.
• 2026: Accepted to CVPR 2026.

---

📦 Release Plan

• Paper link (arXiv)
• Code release
• Pretrained models
• Dataset / Benchmark (if applicable)
• Demo

---

📌 Table of Contents

• Overview
• News
• Release Plan
• Citation
• License
• Contact

---

📚 Citation

If you find this work useful, please consider citing:

@inproceedings{gsv2x_cvpr2026,
  title     = {GSV2X: Geometry-Aware Uncertainty Modeling and Orthogonal Fusion for Robust Roadside Perception},
  author    = {Jianqiang Xu, Gensheng Pei, Huafeng Liu, Yazhou Yao},
  booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year      = {2026}
}

---

📝 License

This project is released under the MIT License.

---

📮 Contact

• GitHub: Kilitoku
• Email: TODO