This repository is a curated list of resources for understanding and working with deepfakes, targeting researchers, developers, and anyone interested in the field of AI-generated synthetic media. It provides a structured overview of papers, code, datasets, and articles related to deepfake generation and detection across video, text, and voice modalities.

How It Works

The list categorizes deepfake materials by modality (video, text, voice) and task (generation, detection). Within each category, it links to seminal research papers, open-source code repositories (e.g., FaceSwap, DeepFaceLab, GLTR), relevant datasets (e.g., FaceForensics++, Celeb-DF), and informative online articles. This organization allows users to quickly find relevant tools, foundational research, and practical examples for deepfake creation and analysis.

Quick Start & Requirements

This is a curated list, not a runnable project. To utilize the resources, users will need to individually clone or download code repositories and set up their respective environments, which typically involve Python and deep learning frameworks like TensorFlow or PyTorch. Specific hardware requirements (e.g., GPUs) will vary per linked project.

Highlighted Details

• Comprehensive coverage of deepfake techniques for video, text, and voice.
• Links to popular open-source deepfake generation tools like FaceSwap and DeepFaceLab.
• Includes resources for deepfake detection, addressing the growing need for forensic analysis.
• Features datasets and challenges crucial for training and evaluating detection models.

Maintenance & Community

The list is maintained by Mengnan Du from Texas A&M University. Contributions are welcomed via pull requests. Contact information for the maintainer is provided via email and Twitter.

Licensing & Compatibility

The licensing of individual linked projects varies. Users must consult the licenses of each specific repository (e.g., MIT, Apache 2.0) for usage restrictions and compatibility with commercial or closed-source applications.

Limitations & Caveats

The maintainer notes that the list is "probably biased and incomplete." While extensive, it may not cover every emerging deepfake resource or technique. The rapid evolution of deepfake technology means some linked resources may become outdated.