This repository is an extensive, curated collection of resources for talking head synthesis, targeting researchers and developers in generative AI, GANs, and NeRFs. It provides a comprehensive overview of papers, code, datasets, tools, and metrics in the field, aiming to facilitate advancements in creating realistic digital human faces.

How It Works

The repository acts as a central hub, aggregating links to academic papers (often with arXiv or direct PDF access), released code repositories, and relevant datasets. It categorizes resources by type (e.g., audio-driven, text-driven, NeRF/3D/Gaussian Splatting) and includes surveys and benchmarks to track progress and compare methodologies. The project actively encourages community contributions to keep the resource up-to-date.

Quick Start & Requirements

This is a curated list of resources, not a runnable software package. Users will need to individually clone, install, and run the code repositories linked within the README. Dependencies vary significantly per project but generally include Python, deep learning frameworks (PyTorch/TensorFlow), and potentially specialized hardware like GPUs with CUDA.

Highlighted Details

• Extensive lists of datasets (e.g., VoxCeleb, CelebV-HQ, HDTF) and papers covering audio-driven, text-driven, and 3D synthesis techniques.
• Detailed sections on metrics (PSNR, SSIM, FID, LPIPS, LSE) and tools/software (LUCIA, Yepic Studio, CrazyTalk).
• Regular updates, with a significant portion of listed papers and projects from 2023-2025, indicating a focus on recent advancements.
• Integration of emerging techniques like 3D Gaussian Splatting and NeRFs alongside traditional GANs.

Maintenance & Community

The repository is actively maintained by Kedreamix, with a call for pull requests and issue submissions to improve the collection. It acknowledges contributions from other curated lists, indicating community engagement.

Licensing & Compatibility

The repository itself is not software and does not have a license. Individual projects linked within will have their own licenses, which users must consult for usage and compatibility.

Limitations & Caveats

As a curated list, the repository does not provide a unified interface or guarantee the functionality or licensing of linked external projects. Users must independently verify the requirements and licenses of each paper, code repository, or dataset they intend to use.