English

SIGMAN:Scaling 3D Human Gaussian Generation with Millions of Assets

Computer Vision and Pattern Recognition 2025-04-10 v1

Abstract

3D human digitization has long been a highly pursued yet challenging task. Existing methods aim to generate high-quality 3D digital humans from single or multiple views, but remain primarily constrained by current paradigms and the scarcity of 3D human assets. Specifically, recent approaches fall into several paradigms: optimization-based and feed-forward (both single-view regression and multi-view generation with reconstruction). However, they are limited by slow speed, low quality, cascade reasoning, and ambiguity in mapping low-dimensional planes to high-dimensional space due to occlusion and invisibility, respectively. Furthermore, existing 3D human assets remain small-scale, insufficient for large-scale training. To address these challenges, we propose a latent space generation paradigm for 3D human digitization, which involves compressing multi-view images into Gaussians via a UV-structured VAE, along with DiT-based conditional generation, we transform the ill-posed low-to-high-dimensional mapping problem into a learnable distribution shift, which also supports end-to-end inference. In addition, we employ the multi-view optimization approach combined with synthetic data to construct the HGS-1M dataset, which contains 11 million 3D Gaussian assets to support the large-scale training. Experimental results demonstrate that our paradigm, powered by large-scale training, produces high-quality 3D human Gaussians with intricate textures, facial details, and loose clothing deformation.

Keywords

Cite

@article{arxiv.2504.06982,
  title  = {SIGMAN:Scaling 3D Human Gaussian Generation with Millions of Assets},
  author = {Yuhang Yang and Fengqi Liu and Yixing Lu and Qin Zhao and Pingyu Wu and Wei Zhai and Ran Yi and Yang Cao and Lizhuang Ma and Zheng-Jun Zha and Junting Dong},
  journal= {arXiv preprint arXiv:2504.06982},
  year   = {2025}
}

Comments

project page:https://yyvhang.github.io/SIGMAN_3D/