We present a novel approach for real-time facial tracking and reconstruction of a monocular video sequence using parametric head models.
conda create -n guided python=3.10 -y
pip install . -e - Specify how to store the FLAME model
Make sure that ninja is installed, e.g. this is needed for the torch.utils.cpp_extension to work.
sudo apt-get install ninja-build
sudo apt-get install libnvidia-gl-535
sudo apt-get install libegl1