Tree-D Fusion: Simulation-Ready Tree Dataset from Single Images with Diffusion Priors

Authors: Jae Joong Lee, Bosheng Li, Sara Beery, Jonathan Huang, Songlin Fei, Raymond A. Yeh, Bedrich Benes

Conference: ECCV 2024 Website: Website

📺 Video

📄 Paper

arxiv link

📊 Dataset

Dataset Links

Updates

• The Tree-D Fusion model is featured in the Venice Biennale 2025 “Re-Leaf” project, developed in collaboration with the MIT Senseable City Lab.

Codes

• Download the ckpt file from here
• Install the dependency packages under "Magic123" and "zero123" folders.
• Run with treedfusion.py --help

BibTeX Citation

If you use Tree-D fusion in a scientific publication, we would appreciate using the following citations:

@inproceedings{lee2024tree,
  title={Tree-D Fusion: Simulation-Ready Tree Dataset from Single Images with Diffusion Priors},
  author={Lee, Jae Joong and Li, Bosheng and Beery, Sara and Huang, Jonathan and Fei, Songlin and Yeh, Raymond A and Benes, Bedrich},
  booktitle={European Conference on Computer Vision},
  pages={439--460},
  year={2024},
  organization={Springer}
}