PyTorch Implementation for Conference Paper "Initiative Defense against Facial Manipulation (AAAI 2021)"

This repository provides the official PyTorch implementation of the following paper:

Initiative Defense against Facial Manipulation (AAAI 2021)
https://ojs.aaai.org/index.php/AAAI/article/view/16254/16061

Preparation

Download pretrained models from pretrained_model and put them into ./checkpoints.

Download clean faces for test from clean_faces and unzip them into ./clean_faces.

Also, you can download the whole CeleBA dataset for test by

bash download.sh celeba

Usage

For test, you can directly run the following commands:

# Test with the noise generator defense
python main.py --mode test --dataset CelebA --image_size 128 \
               --c_dim 5 --g_repeat_num 9 --batch_size 1 \
               --selected_attrs Black_Hair Gray_Hair Pale_Skin No_Beard Eyeglasses \
               --celeba_image_dir ./clean_faces --eps 0.03

# Test without the noise generator defense
python main.py --mode test --dataset CelebA --image_size 128 \
               --c_dim 5 --g_repeat_num 9 --batch_size 1 \
               --selected_attrs Black_Hair Gray_Hair Pale_Skin No_Beard Eyeglasses \
               --celeba_image_dir ./clean_faces --eps 0.03 --use_PG False

Citation

If you find this work useful for your research, please cite our paper:

@inproceedings{huang2021initiative,
author={Qidong Huang and Jie Zhang and Wenbo Zhou and Weiming Zhang and Nenghai Yu},
title={Initiative Defense against Facial Manipulation},
booktitle={Proceedings of the AAAI Conference on Artificial Intelligence (AAAI)},
year={2021}
}

License

The code is released under MIT License (see LICENSE file for details).

Acknowledgements

This work is heavily based on StarGAN.