Dataset for "Synthetic data augmentation for facial re-identification."

A dataset called syn-Face consisting of ~1,800,000 images of syntethically generated faces. Images were constructed by generating ~4000 3D humanoid models, and then rendering 2D images of them in 4 different lighting scenarions, 28 different camera positions, and 4 different resolutions. See "Synthetic data augmentation for facial re-identification", G. Brown et al, 2019 for more information. The file uploaded here contains a representive sample: for access to the full dataset please contact the creator.