Abstract

Existing methods like Neural Radiation Fields (NeRF) and 3D Gaussian Splatting (3DGS) have made significant strides in facial attribute control such as facial animation and components editing, yet they struggle with fine-grained representation and scalability in dynamic head modeling. To address these limitations, we propose GaussianStyle, a novel framework that integrates the volumetric strengths of 3DGS with the powerful implicit representation of StyleGAN. The GaussianStyle preserves structural information, such as expressions and poses, using Gaussian points, while projecting the implicit volumetric representation into StyleGAN to capture high-frequency details and mitigate the over-smoothing commonly observed in neural texture rendering. Experimental outcomes indicate that our method achieves state-of-the-art performance in reenactment, novel view synthesis, and animation.

Method

Figure. 2. Overview of GaussianStyle. The proposed Tri-Stage training strategy includes StyleGAN-based Volumetric Rendering. In Stage 1, we construct static coarse canonical Gaussians. In Stage 2, Gaussians are queried from a temporal-aware triplane for attention-based deformation. In Stage 3, we initialize the StyleGAN through multi-view PTI initialization and project dynamic Gaussian prior into StyleGAN for volumetric rendering.

Demo Video

We have compressed the videos and downgrade quality to accommodate the 200MB supplementary material submission limition.

BibTeX

@misc{liu2024gaussianstylegaussianheadavatar,
      title={GaussianStyle: Gaussian Head Avatar via StyleGAN}, 
      author={Pinxin Liu and Luchuan Song and Daoan Zhang and Hang Hua and Yunlong Tang and Huaijin Tu and Jiebo Luo and Chenliang Xu},
      year={2024},
      eprint={2402.00827},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2402.00827}, 
}