4DFAB is a large-scale dynamic high-resolution 3D face dataset consisting of 1.8+ million 3D meshes from 180 participants (collected from 4 different conferences over 5 years, 2012-2017) whose 4D videos show both spontaneous and posed facial expressions.

This dataset can be used for face recognition and facial expression recognition, as well as behavioral biometrics. It can also be used to learn blendshape and parameterize facial behavior.