UniMamba

Unified Spatial-Channel Representation Learning with Group-Efficient Mamba for LiDAR-based 3D Object Detection

Xin Jin^♠️,2,3, Haisheng Su^{♠️,1,3 📧}, Kai Liu³, Cong Ma³, Wei Wu^3,4, Fei HUI^{3 📧}, Junchi Yan^{1 📧}

¹ School of Computer Science, Shanghai Jiao Tong University

² Chang'an University, ³ SenseAuto Research, ⁴ Tsinghua University

^♠️ Equal Contributions, ^📧 Corresponding authors

• Mar. 9th, 2025: We released our paper on Arxiv. Code/Models are coming soon. Please stay tuned! ☕️
• Mar. 9th, 2025: Our paper has been accepted to CVPR 2025!

• Introduction
• Framework
• Evaluation on nuScenes dataset
• Evaluation on Waymo Open dataset
• Evaluation on Argoverse2 dataset
• License
• Contact
• Citation

TBD

This project is released under the MIT license

If you have any questions, please contact Haisheng Su via email ([email protected]).

If you find UniMamba is useful in your research or applications, please consider giving us a star 🌟 and citing it by the following BibTeX entry.

@article{su2025unimamba,
  title={UniMamba: Unified Spatial-Channel Representation Learning with Group-Efficient Mamba for LiDAR-based 3D Object Detection},
  author={Jin, Xin and Su, Haisheng and Liu, Kai and Ma, Cong and Wu, Wei and Hui, Fei and Yan, Junchi},
  journal={arXiv preprint arXiv:2503.12009},
  year={2025}
}