Official Code for Paper ``Improving Adversarial Robustness of 3D Point Cloud Classification Models''(ECCV 2022)
- Tensorflow>=1.14.0 (not support Tensorflow 2.0)
- Pytorch>=1.2.0
- PointCutMix-K
Please follow this repo.
The ModelNet40 can be downloaded from here.
Pretrained model can be found in here. Or, you can train your own model from scratch.
python train.py
Parts of code are from DGCNN, PointCloud-Saliency-Map and PointCutMix-K.
If you find our work is useful, please cite it with the following format:
@inproceedings{li2022improving,
title={Improving Adversarial Robustness of 3D Point Cloud Classification Models},
author={Li, Guanlin and Xu, Guowen and Qiu, Han and He, Ruan and Li, Jiwei and Zhang, Tianwei},
booktitle={European Conference on Computer Vision},
pages={672--689},
year={2022},
organization={Springer}
}
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent