3DGH: 3D Head Generation with Composable Hair and Face

TL;DR

3DGH introduces a novel 3D generative model for human heads that separately models hair and face components, enabling realistic synthesis and hairstyle editing through a dual-generator architecture and a new data representation.

Contribution

It proposes a new data representation with template-based 3D Gaussian Splatting and a dual-GAN architecture with cross-attention for separate yet correlated hair and face modeling.

Findings

Outperforms state-of-the-art 3D GANs in head synthesis quality.

Enables realistic hairstyle editing and composable head generation.

Demonstrates stable training with synthetic renderings.

Abstract

We present 3DGH, an unconditional generative model for 3D human heads with composable hair and face components. Unlike previous work that entangles the modeling of hair and face, we propose to separate them using a novel data representation with template-based 3D Gaussian Splatting, in which deformable hair geometry is introduced to capture the geometric variations across different hairstyles. Based on this data representation, we design a 3D GAN-based architecture with dual generators and employ a cross-attention mechanism to model the inherent correlation between hair and face. The model is trained on synthetic renderings using carefully designed objectives to stabilize training and facilitate hair-face separation. We conduct extensive experiments to validate the design choice of 3DGH, and evaluate it both qualitatively and quantitatively by comparing with several state-of-the-art 3D GAN methods, demonstrating its effectiveness in unconditional full-head image synthesis and composable 3D hairstyle editing. More details will be available on our project page: https://c-he.github.io/projects/3dgh/.