The official code of the paper "Standing on the Shoulders of Giants: Reprogramming Visual-Language Model for General Deepfake Detection"
This repository currently provides a preview version of the source code. Comprehensive usage guidelines will be released in upcoming updates.
For face embedding extraction, we adopt the pre-trained face model released by the TransFace project, available at the link below: Google Drive
We gratefully acknowledge the following repositories, which our implementation builds upon:
If the code can be helpful for you, please cite our papers as follows:
@inproceedings{lin2025standing,
title={Standing on the shoulders of giants: Reprogramming visual-language model for general deepfake detection},
author={Lin, Kaiqing and Lin, Yuzhen and Li, Weixiang and Yao, Taiping and Li, Bin},
booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
volume={39},
number={5},
pages={5262--5270},
year={2025}
}