AmodalGen3D: Generative Amodal 3D Object Reconstruction from Sparse Unposed Views

Junwei Zhou, Yu-Wing Tai

Dartmouth College

TL;DR: AmodalGen3D intergrates 2D amodal completion and 3D generative modeling to achieve amodal 3D objects reconstruction from sparse unposed views that are even heavily occluded.

Official implementation for paper 'AmodalGen3D: Generative Amodal 3D Object Reconstruction from Sparse Unposed Views'.

Intergrating 2D amodal completion prior and 3D generative modeling ability in the latent 3D space to achieve amodal 3D objects reconstruction from sparse unposed views, under various scenarios, including single object-level, in-the-wild and in-the-scene.

Updates and TODOs

• ✔️ Update@12/04/2025: initialize the project page.
• 🔲 TODO: The code will be released once accepted. Please stay tuned!

Citation

Here is the bibtex reference. If you find our work interesting or useful, please give us a ⭐ or cite our paper.

@misc{zhou2025amodalgen3d,
      title={AmodalGen3D: Generative Amodal 3D Object Reconstruction from Sparse Unposed Views}, 
      author={Junwei Zhou and Yu-Wing Tai},
      year={2025},
      eprint={2511.21945},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2511.21945}, 
}