LEGO: A Lifting-Free Approach for Exocentric-to-Egocentric Video Generation

TL;DR Given a single third-person video, LEGO generates what a person in it sees, without estimating depth or building a point cloud.

• 📄 Paper: arXiv:2610.12442
• 🌐 Project page: https://suhwan-cho.github.io/LEGO/
• 💻 Code: https://github.com/suhwan-cho/LEGO

Model weights

Coming soon. We are polishing the code and checkpoints and will release them here.

Citation

@article{cho2026lego,
  title   = {LEGO: A Lifting-Free Approach for Exocentric-to-Egocentric Video Generation},
  author  = {Cho, Suhwan and Choi, Yonwoo and Kim, Soongjin and Park, Jicheol and Lim, Taegyu},
  journal = {arXiv preprint arXiv:2610.12442},
  year    = {2026}
}