Awesome

VideoFACT (WACV 2024): Detecting Video Forgeries Using Attention, Scene Context, and Forensic Traces

[Paper] [Supplemental] [Poster] [Presentation]

News

• 2024.01.06: Our code for inference is released.

Overview

Work in Progress

• Dataset Publication
• Training Code
• Evaluation Code

Dependencies and Installation

1. Clone Repo

   git clone https://github.com/ductai199x/videofact-wacv-2024.git
   

2. Create Virtual Environment and Install Dependencies

   With virtualenv:

   virtualenv .venv --python=python3.9
   source .venv/bin/activate
   

   With conda:

   conda create -n videofact python=3.9
   conda activate videofact
   

   Install dependencies:

   pip install -r requirements.txt
   

Get Started

Prepare pretrained models

Before performing the following steps, please download our pretrained models first.

[VideoFACT Xfer] [VideoFACT Deepfake]

Run the demo

python demo.py

Citation

If you find our work useful in your research, please consider citing:

@inproceedings{nguyen2024videofact,
  title={VideoFACT: Detecting video forgeries using attention, scene context, and forensic traces},
  author={Nguyen, Tai D and Fang, Shengbang and Stamm, Matthew C},
  booktitle={Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision},
  pages={8563--8573},
  year={2024}
}

License

Licensed under a Creative Commons Attribution-NonCommercial 4.0 International for Non-commercial use only. Any commercial use should get formal permission first.