This is the official repository for our paper DivTrackee versus DynTracker: Promoting Diversity in Anti-Facial Recognition against Dynamic FR Strategy.

• Build environment

# use anaconda to build environment 
conda create -n div2trackee python=3.8
conda activate div2trackee
# install packages
pip install -r requirements.txt

Quick Start

1. Place the pre-trained StyleGAN2 weights in the 'pretrained_models' folder.

2. Download face recognition models at here and place them at 'models' folder.

3. Using e4e method to get the latent codes of facial images.

4. Run main.py.

    python main.py --data_dir input_images --noise_path noises.pt --latent_path latents.pt --checkpoint_dir checkpoint_dir --output_dir output

5. Run eval.py.

Citation

If the code and paper help your research, please kindly cite:

@article{fan2025divtrackee,
  title={DivTrackee versus DynTracker: Promoting Diversity in Anti-Facial Recognition against Dynamic FR Strategy},
  author={Fan, Wenshu and Zhang, Minxing and Li, Hongwei and Jiang, Wenbo and Chen, Hanxiao and Yue, Xiangyu and Backes, Michael and Zhang, Xiao},
  booktitle={Proceedings of the ACM SIGSAC Conference on Computer and Communications Security},
  year={2025}
}