Exploring Disentangled Feature Representation Beyond Face Identification

Abstract

This paper proposes learning disentangled but complementary face features with minimal supervision by face identification. Specifically, we construct an identity Distilling and Dispelling Autoencoder (D2AE) framework that adversarially learns the identity-distilled features for identity verification and the identity-dispelled features to fool the verification system. Thanks to the design of two-stream cues, the learned disentangled features represent not only the identity or attribute but the complete input image. Comprehensive evaluations further demonstrate that the proposed features not only maintain state-of-the-art identity verification performance on LFW, but also acquire competitive discriminative power for face attribute recognition on CelebA and LFWA. Moreover, the proposed system is ready to semantically control the face generation/editing based on various identities and attributes in an unsupervised manner.

arXiv commentAccepted by CVPR 2018

References in corpus (2)

• MS-Celeb-1M: A Dataset and Benchmark for Large-Scale Face Recognition
• Beyond Face Rotation: Global and Local Perception GAN for Photorealistic and Identity Preserving Frontal View Synthesis

Cited by in corpus (4)

• Teacher Supervises Students How to Learn From Partially Labeled Images for Facial Landmark Detection
• From Synthetic to Real: Image Dehazing Collaborating with Unlabeled Real Data
• Disentanglement for Discriminative Visual Recognition
• Beyond Trade-off: Accelerate FCN-based Face Detector with Higher Accuracy