GART: Gaussian Articulated Template Models

Jiahui Lei, Yufu Wang, Georgios Pavlakos, Lingjie Liu, Kostas Daniilidis

CVPR 2024 pp. 19876-19887

doi:10.1109/CVPR52733.2024.01879 /cvpr/2024/lei2024cvpr-gart/

Abstract

We introduce Gaussian Articulated Template Model (GART) an explicit efficient and expressive representation for non-rigid articulated subject capturing and rendering from monocular videos. GART utilizes a mixture of moving 3D Gaussians to explicitly approximate a deformable subject's geometry and appearance. It takes advantage of a categorical template model prior (SMPL SMAL etc.) with learnable forward skinning while further generalizing to more complex non-rigid deformations with novel latent bones. GART can be reconstructed via differentiable rendering from monocular videos in seconds or minutes and rendered in novel poses faster than 150fps.

PDF CVPR Semantic Scholar

Cite

Text

Lei et al. "GART: Gaussian Articulated Template Models." Conference on Computer Vision and Pattern Recognition, 2024. doi:10.1109/CVPR52733.2024.01879

Markdown

[Lei et al. "GART: Gaussian Articulated Template Models." Conference on Computer Vision and Pattern Recognition, 2024.](https://mlanthology.org/cvpr/2024/lei2024cvpr-gart/) doi:10.1109/CVPR52733.2024.01879

BibTeX

@inproceedings{lei2024cvpr-gart,
  title     = {{GART: Gaussian Articulated Template Models}},
  author    = {Lei, Jiahui and Wang, Yufu and Pavlakos, Georgios and Liu, Lingjie and Daniilidis, Kostas},
  booktitle = {Conference on Computer Vision and Pattern Recognition},
  year      = {2024},
  pages     = {19876-19887},
  doi       = {10.1109/CVPR52733.2024.01879},
  url       = {https://mlanthology.org/cvpr/2024/lei2024cvpr-gart/}
}