Skip to content

Repository files navigation

LOD_PDR

Official repository of "LiDAR-based All-weather 3D Object Detection via Prompting and Distilling 4D Radar", ECCV2024. [Paper]

Requirements

The code has been tested with

  • python 3.8
  • CUDA 11.1
  • pytorch 1.10.1
  • spconv-cu111 2.1.25
  • open3d 0.15.2
  • opencv-python 4.8.1.78
  • matplotlib 3.5.3
  • numba 0.53.0
  • nms 0.1.6

Usage

Download logs

https://drive.google.com/drive/folders/13JRwnTY_cupN13-gwpXWlfIsLmIAIaRZ?usp=sharing

Teacher

# Train
python main_train_0.py

# Test
python main_cond_0.py

Student

# Train
python main_train_student.py

# Test
python main_cond_student.py

Citation

If you find our work helpful, please consider citing our paper:

@inproceedings{Chae_2024_ECCV,
  title={LiDAR-Based All-Weather 3D Object Detection via Prompting and Distilling 4D Radar},
  author={Chae, Yujeong and Kim, Hyeonseong and Oh, Changgyoon and Kim, Minseok and Yoon, Kuk-Jin},
  booktitle={European Conference on Computer Vision},
  pages={368--385},
  year={2024},
  organization={Springer}
}

Acknowledgements

This work is developed based on the K-Radar dataset and codebase.

About

Official repository of "LiDAR-based All-weather 3D Object Detection via Prompting and Distilling 4D Radar" (ECCV2024)

Resources

Stars

13 stars

Watchers

2 watching

Forks

Releases

Packages

Contributors

Languages