Timezone: »

Dual-Agent GANs for Photorealistic and Identity Preserving Profile Face Synthesis
Jian Zhao · Lin Xiong · Panasonic Karlekar Jayashree · Jianshu Li · Fang Zhao · Zhecan Wang · Panasonic Sugiri Pranata · Panasonic Shengmei Shen · Shuicheng Yan · Jiashi Feng

Mon Dec 04 06:30 PM -- 10:30 PM (PST) @ Pacific Ballroom #131

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. To narrow this gap, we propose a Dual-Agent Generative Adversarial Network (DA-GAN) model, which can improve the realism of a face simulator's output using unlabeled real faces, while preserving the identity information during the realism refinement. The dual agents are specifically designed for distinguishing real v.s. fake and identities simultaneously. In particular, we employ an off-the-shelf 3D face model as a simulator to generate profile face images with varying poses. DA-GAN leverages a fully convolutional network as the generator to generate high-resolution images and an auto-encoder as the discriminator with the dual agents. Besides the novel architecture, we make several key modifications to the standard GAN to preserve pose and texture, preserve identity and stabilize training process: (i) a pose perception loss; (ii) an identity perception loss; (iii) an adversarial loss with a boundary equilibrium regularization term. Experimental results show that DA-GAN not only presents compelling perceptual results but also significantly outperforms state-of-the-arts on the large-scale and challenging NIST IJB-A unconstrained face recognition benchmark. In addition, the proposed DA-GAN is also promising as a new approach for solving generic transfer learning problems more effectively.

Author Information

Jian Zhao (National University of Singapore)

聚焦 Knowledge changes fate.

Lin Xiong (Panasonic R&D Center Singapore)

Lin Xiong received the B.S. degree from Shaanxi University of Science & Technology in 2003, and he received the Ph.D. degree with School of Electronic Engineering, Xidian University, China, in 2014. He is currently a research engineer of Learning & Vision, Core Technology Group, Panasonic R&D Center Singapore, Singapore. His current research interests include unconstrained/large-scale face recognition, person re-identification, deep learning architecture engineering, transfer learning, Riemannian manifold optimization, sparse and low-rank matrix factorization.

Panasonic Karlekar Jayashree (Panasonic, Singapore)
Jianshu Li (National University of Singapore)
Fang Zhao (National University of Singapore)
Zhecan Wang (Franklin. W. Olin College of Engineering)
Panasonic Sugiri Pranata (Panasonic, Singapore)
Panasonic Shengmei Shen (Panasonic, Singapore)
Shuicheng Yan (National University of Singapore)
Jiashi Feng (National University of Singapore)

More from the Same Authors