This repository demonstrates

1. how to create a FLAME texture model from the BFM vertex color space, and
2. how to convert a BFM mesh to a FLAME mesh.

FLAME is a lightweight and expressive generic head model learned from over 33,000 of accurately aligned 3D scans. For a list of public FLAME resources (publications, code repositories, datasets), please visit the 🔥 FLAME-Universe 🔥.

Install pip and virtualenv

sudo apt-get install python3-pip python3-venv

Clone the git project:

git clone https://github.com/TimoBolkart/BFM_to_FLAME.git

Set up and activate virtual environment:

mkdir <your_home_dir>/.virtualenvs
python3 -m venv <your_home_dir>/.virtualenvs/BFM_to_FLAME
source <your_home_dir>/.virtualenvs/BFM_to_FLAME/bin/activate

Make sure your pip version is up-to-date:

pip install -U pip

Install requirements

pip install numpy==1.19.4
pip install h5py==3.1.0
pip install chumpy==0.70 
pip install opencv-python==4.4.0.46

Download BFM 2017 (i.e. 'model2017-1_bfm_nomouth.h5') from here and place it in the model folder. Download inpainting masks from here and place it in the data folder.

Running

python col_to_tex.py

outputs a 'FLAME_albedo_from_BFM.npz' in the output folder. This file can be used with several FLAME-based repositories like TF_FLAME or FLAME photometric optimization.

Install mesh processing libraries from MPI-IS/mesh within the virtual environment. Download FLAME from here and place it in the model folder.

Running

python mesh_convert.py

outputs a FLAME mesh for a specified BFM mesh. The demo supports meshes in 'BFM 2017', 'BFM 2009', or 'cropped BFM 2009' (i.e. as used by 3DDFA) topology.

When using this code, the generated texture space, or FLAME meshes in a scientific publication, please cite

@article{FLAME:SiggraphAsia2017,
  title = {Learning a model of facial shape and expression from {4D} scans},
  author = {Li, Tianye and Bolkart, Timo and Black, Michael. J. and Li, Hao and Romero, Javier},
  journal = {ACM Transactions on Graphics, (Proc. SIGGRAPH Asia)},
  volume = {36},
  number = {6},
  year = {2017},
  url = {https://doi.org/10.1145/3130800.3130813}
}

When using the converted texture space, please further follow the license agreement of the BFM model as specified here.

We thank the authors of the BFM 2017 model for making the model publicly available.