Fig. 8From: Deep-learned faces: a surveySample data in three benchmarks LFW (top), video frames from IJB-A (middle) and MegaFace (bottom)Back to article page