If you use this codebase in your experiments please cite:

[Face Normals “in-the-wild” using Fully Convolutional Networks G. Trigeorgis, P. Snape, I. Kokkinos, S. Zafeiriou. IEEE Conference on Computer Vision and Pattern Recognition (CVPR). July 2017. ]

We highly recommended to use conda as your Python distribution. Once downloading and installing conda, this project can be installed by:

Step 1: Create a new conda environment and activate it:

$ conda create -n normals python=3.5
$ source activate normals

Step 1: Install TensorFlow v1 following the official installation instructions. For example for the gpu enabled version of TensorFlow

(normals)$ pip install --upgrade tensorflow-gpu

Step 2: Install Menpo following the official installation instructions. This is not mandatory, but makes importing and visualisation of images a breeze.

(normals)$ conda install -c menpo menpo

Disclaimer: The pretrained models can only be used for non-commercial academic purposes.