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

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

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

Computer Vision and Pattern Recognition 2024-07-16 v1

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

Abstract

We introduce Tree D-fusion, featuring the first collection of 600,000 environmentally aware, 3D simulation-ready tree models generated through Diffusion priors. Each reconstructed 3D tree model corresponds to an image from Google's Auto Arborist Dataset, comprising street view images and associated genus labels of trees across North America. Our method distills the scores of two tree-adapted diffusion models by utilizing text prompts to specify a tree genus, thus facilitating shape reconstruction. This process involves reconstructing a 3D tree envelope filled with point markers, which are subsequently utilized to estimate the tree's branching structure using the space colonization algorithm conditioned on a specified genus.

Keywords

text-to-3d generation diffusion model random forest

Cite

@article{arxiv.2407.10330,
  title  = {Tree-D Fusion: Simulation-Ready Tree Dataset from Single Images with Diffusion Priors},
  author = {Jae Joong Lee and Bosheng Li and Sara Beery and Jonathan Huang and Songlin Fei and Raymond A. Yeh and Bedrich Benes},
  journal= {arXiv preprint arXiv:2407.10330},
  year   = {2024}
}

Comments

Accepted to ECCV24

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

Abstract

Keywords

Cite

Comments

Related papers