memoavatar/memo_data

MEMO Video Dataset

This dataset contains a curated collection of human talking videos gathered from publicly accessible sources on the internet. It is part of the training data for the MEMO model. Before using this dataset, please read the following guidelines carefully to ensure compliance with legal, ethical, and licensing requirements. By using this dataset, you acknowledge that you have read, understood, and agree to abide by the following terms.

Dataset Overview

• Contents: A set of MP4 video files showing individuals speaking. No personally identifying metadata is included beyond the video content itself.
• Source: Publicly available videos (e.g., Creative Commons, public domain, or videos with explicit permissions) from various online platforms.
• Purpose: Research and development of talking-video generation models, specifically as part of the MEMO training pipeline.

Directory Structure

All files are stored in the root of the directory and are named arbitrarily.

Legal & Copyright Notice

1. Source Compliance: All videos included in this dataset were sourced only from materials that are clearly in the public domain.
2. No Unlicensed Content: If you identify any video in this dataset that you believe is subject to copyright and was not released under a compatible license or permission, please contact us immediately to have it removed.
3. Research Use Only: This dataset is licensed strictly for non-commercial, research, and educational purposes. Any commercial use, redistribution, or public display beyond the scope of research requires separate permission from the respective rights holders.
4. No Model Endorsement: Use of this dataset to generate synthetic videos of individuals should avoid creating misleading, defamatory, or malicious content (e.g., deepfakes for impersonation or misinformation).

Citation

If you use this dataset in your work, please cite the MEMO paper:

@article{zheng2024memo,
  title={MEMO: Memory-Guided Diffusion for Expressive Talking Video Generation},
  author={Zheng, Longtao and Zhang, Yifan and Guo, Hanzhong and Pan, Jiachun and Tan, Zhenxiong and Lu, Jiahao and Tang, Chuanxin and An, Bo and Yan, Shuicheng},
  journal={arXiv preprint arXiv:2412.04448},
  year={2024}
}