README.md

SwapAnyone: Consistent and Realistic Video Synthesis for Swapping Any Person into Any Video

If you like our project, please give us a star ⭐ on GitHub for latest update.

     

You can visit our project page to get the video results.

Results of SwapAnyone.

Implemetation of SwapAnyone: Consistent and Realistic Video Synthesis for Swapping Any Person into Any Video

🗓️ TODO

We will update the following list after the paper is accepted.

• [2025-03-13] We have released our project page.
• We have uploaded our paper, SwapAnyone on arXiv
• Upload the code
• Open-source training
• Open-source datasets

🌅 Comparisons with others.

📊 Quantitative comparison

📊 Human evaluation test statistics

🤝 Contributors

🙏 Acknowledgements

• AnimateDiff - https://github.com/guoyww/animatediff/
• AnimateAnyone(official) - https://arxiv.org/pdf/2311.17117
• AnimateAnyone(Moore-AnimateAnyone) - https://github.com/MooreThreads/Moore-AnimateAnyone
• Stable Diffusion - https://github.com/CompVis/stable-diffusion

BibTeX

@misc{SwapAnyone,
        title={SwapAnyone: Consistent and Realistic Video Synthesis for Swapping Any Person into Any Video}, 
        author={Chengshu Zhao and Yunyang Ge and Xinhua Cheng and Bin Zhu and Yatian Pang and Bin Lin and Fan Yang and Feng Gao and Li Yuan},
        year={2025},
        eprint={2503.09154},
        archivePrefix={arXiv},
        primaryClass={cs.CV},
        url={https://arxiv.org/abs/2503.09154}, 
    }