🌟 News: Our paper "Is Discretization Fusion All You Need for Collaborative Perception?" has been accepted to ICRA 2025!

• 📦 Is Discretization Fusion All You Need for Collaborative Perception?
  • 📚 Table of Contents
  • 🧩 Code
  • 💻 Installation
  • 🚀 Getting Started
    • 📥 Download Datasets
    • 🏃‍♂️ Train
    • 🧪 Evaluation
  • 📖 Citation
  • 🙏 Acknowledgements

Code is available in this repository. Please refer to individual scripts and folders for usage details.

Please refer to the detailed documentation in docs/INSTALL.md.

You can also install required packages via pip:

pip install -r requirements.txt

However, this is likely to result in package dependency errors.

• OPV2V: https://mobility-lab.seas.ucla.edu/opv2v/
• DAIR-V2X: https://github.com/AIR-THU/DAIR-V2X?tab=readme-ov-file#dataset

To start training, simply run:

./train.sh

To run evaluation (inference):

./inference.sh

If you use this work, please consider citing our paper.

@article{yang2025discretization,
  title={Is Discretization Fusion All You Need for Collaborative Perception?},
  author={Yang, Kang and Bu, Tianci and Li, Lantao and Li, Chunxu and Wang, Yongcai and Li, Deying},
  journal={arXiv preprint arXiv:2503.13946},
  year={2025}
}

• Thanks to the authors of OpenCOOD 、 Sparse4D and Where2comm for providing excellent codebases.
• We also appreciate the datasets provided by DAIR-V2X and OPV2V.