GRACE: Graph-Regularized Attentive Convolutional Entanglement with Laplacian Smoothing for Robust DeepFake Video Detection

The official pytorch implementation of "GRACE: Graph-Regularized Attentive Convolutional Entanglement with Laplacian Smoothing for Robust DeepFake Video Detection". Submitted to IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI 2024).

[Paper Link (arXiv)]

Chih-Chung Hsu, Shao-Ning Chen, Mei-Hsuan Wu, Yi-Fang Wang, Chia-Ming Lee, Yi-Shiuan Chou

Advanced Computer Vision LAB, National Cheng Kung University

Overview

In the field of Deepfake detection, one particular issue lies with facial images being mis-detected, often originating from camera motion, degraded videos or adversarial attacks, leading to unexpected temporal artifacts that can undermine the efficacy of DeepFake video detection techniques.

This paper introduces a novel method for robust DeepFake video detection, harnessing the power of the proposed Graph-Regularized Attentive Convolutional Entanglement (GRACE) based on the graph convolutional network with graph Laplacian to address the aforementioned challenges.

Todo List

• 💪 GRACE is already under major revision.
• Release model and training scripts.

Environment

• PyTorch >= 1.7
• CUDA >= 11.2
• python==3.8.18
• pytorch==1.11.0
• cudatoolkit=11.3
• onnx==1.14.1
• onnxruntime==1.16.1

Installation

git clone https://github.com/ming053l/GRACE.git
conda create --name grace python=3.8 -y
conda activate grace
# CUDA 11.3
conda install pytorch==1.12.1 torchvision==0.13.1 torchaudio==0.12.1 cudatoolkit=11.3 -c pytorch
cd GRACE
pip install -r requirements.txt

How to Train

coming soon...

How to Test

coming soon...

Citations

If our work is helpful to your reaearch, please kindly cite our work. Thank!

BibTeX

@misc{hsu2024gracegraphregularizedattentiveconvolutional,
      title={GRACE: Graph-Regularized Attentive Convolutional Entanglement with Laplacian Smoothing for Robust DeepFake Video Detection}, 
      author={Chih-Chung Hsu and Shao-Ning Chen and Mei-Hsuan Wu and Yi-Fang Wang and Chia-Ming Lee and Yi-Shiuan Chou},
      year={2024},
      eprint={2406.19941},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2406.19941}, 
}

Contact

If you have any question, please email zuw408421476@gmail.com to discuss with the author.