1 paper
Gaojian Wang, Feng Lin, Tong Wu +3
This work asks: with abundant, unlabeled real faces, how to learn a robust and transferable facial representation that boosts various face security tasks with respect to generaliza…