Awesome

This project hosts the official implementation for the paper:

Deep Dive into Gradients: Better Optimization for 3D Object Detection with Gradient-Corrected IoU Supervision [PDF][BibTex]

( accepted by CVPR 2023).

Setup

pip install spconv-cu111
pip install Cmake
pip install -r requirement.txt
pip install mayavi
python setup.py develop

cd pcdet/ops/iou3d/cuda_op
python setup.py install

Training

Data Prepare Download KITTI and organize it into the following form:

├── data
│   ├── kitti
│   │   │── ImageSets
│   │   │── training
│   │   │   ├──calib & velodyne & label_2 & image_2 & (optional: planes)
│   │   │── testing
│   │   │   ├──calib & velodyne & image_2

Generatedata infos: python -m pcdet.datasets.kitti.kitti_dataset create_kitti_infos tools/cfgs/dataset_configs/kitti_dataset.yaml
Creat .yaml file
Training the model via python tools/train.py

Visualizations

demo

Citation

If you find our work or code useful in your research, please consider citing:

@inproceedings{ming2023deep,
  title={Deep Dive Into Gradients: Better Optimization for 3D Object Detection With Gradient-Corrected IoU Supervision},
  author={Ming, Qi and Miao, Lingjuan and Ma, Zhe and Zhao, Lin and Zhou, Zhiqiang and Huang, Xuhui and Chen, Yuanpei and Guo, Yufei},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  pages={5136--5145},
  year={2023}
}

Feel free to contact me if you have any questions.