StyleMM: Stylized 3D Morphable Face Model via Text-Driven Aligned Image Translation (PG 2025 Journal Track)

[PG2025] Official repository of EAS of StyleMM [Project Page]

This repository is code for EAS, not full Pipeline. Code for full pipeline is StyleMM Code

Getting Started

• install dependency

bash set.sh

• Put EAM weight and image folders to change the style

EAM Weight and Realistic Facial Images

How to inference Scripts

python inference.py --num_styles 3

To change style, check align_utils.py

Acknowledgments

our codes were based on Diffusers library, ControlNet, and SDXL.

If found EAS or StyleMM useful for your research, please cite our paper:

@article{lee2025stylemm,
  title={StyleMM: Stylized 3D Morphable Face Model via Text-Driven Aligned Image Translation},
  author={Lee, Seungmi and Yun, Kwan and Noh, Junyong},
  journal={arXiv preprint arXiv:2508.11203},
  year={2025}
}