Dual-Agent GANs for Photorealistic and Identity Preserving Profile Face Synthesis

NeurIPS 2017 Jian ZhaoLin XiongPanasonic Karlekar JayashreeJianshu LiFang ZhaoZhecan WangPanasonic Sugiri PranataPanasonic Shengmei ShenShuicheng YanJiashi Feng

Synthesizing realistic profile faces is promising for more efficiently training deep pose-invariant models for large-scale unconstrained face recognition, by populating samples with extreme poses and avoiding tedious annotations. However, learning from synthetic faces may not achieve the desired performance due to the discrepancy between distributions of the synthetic and real face images... (read more)

PDF Abstract

Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK BENCHMARK
Face Verification IJB-A Dual-Agent GANs TAR @ FAR=0.01 97.60% # 1

Methods used in the Paper