Exploring Disentangled Feature Representation Beyond Face Identification

@article{Liu2018ExploringDF,
  title={Exploring Disentangled Feature Representation Beyond Face Identification},
  author={Yu Liu and Fangyin Wei and Jing Shao and Lu Sheng and Junjie Yan and Xiaogang Wang},
  journal={2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year={2018},
  pages={2080-2089}
}
  • Yu Liu, Fangyin Wei, Xiaogang Wang
  • Published 10 April 2018
  • Computer Science
  • 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition
This paper proposes learning disentangled but complementary face features with a minimal supervision by face identification. [] Key Method 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 preserve state-of-the-art identity verification performance on LFW, but also acquire comparable discriminative power for face attribute…
Robust Discrimination and Generation of Faces using Compact, Disentangled Embeddings
TLDR
This work presents an architecture that learns an unsupervised, low-dimensional embedding of faces using an adversarial autoencoder that is able to synthesize high-quality face images and applies the architecture's robust image synthesis capabilities to visually debug label-quality issues in an existing face dataset.
Disentanglement for Discriminative Visual Recognition
TLDR
This chapter systematically summarize the popular and practical solution for disentanglement to achieve more discriminative visual recognition and achieves top performance on a serial of tasks.
Cross-VAE: Towards Disentangling Expression from Identity For Human Faces
TLDR
This paper proposes to extend conditional VAE to a crossed version named Cross-VAE, which is able to use partially labeled data to disentangle expression from identity, and utilizes a symmetric training procedure where the output of each encoder is fed as the condition of the other.
Face identity disentanglement via latent space mapping
TLDR
This paper presents a method that learns how to represent data in a disentangled way, with minimal supervision, manifested solely using available pre-trained networks, by employing a leading pre- trained unconditional image generator, such as StyleGAN.
Feature-Level Frankenstein: Eliminating Variations for Discriminative Recognition
  • Xiaofeng Liu, Site Li, B. Kumar
  • Computer Science
    2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
  • 2019
TLDR
This paper proposes equipping an end-to-end conditional adversarial network with the ability to decompose an input sample into three complementary parts and achieves top performance on a serial of tasks, including digits recognition, lighting, makeup, disguise-tolerant face recognition, and facial attributes recognition.
Learning Disentangled Representation for Robust Person Re-identification
TLDR
A new generative adversarial network, dubbed identity shuffle GAN (IS-GAN), is introduced that factorizes identity-related and -unrelated features from person images using identification labels without any auxiliary information and proposes an identity shuffling technique to regularize the disentangled features.
Powerful Speaker Embedding Training Framework by Adversarially Disentangled Identity Representation
TLDR
Experiments prove that the framework proposed can significantly improve the performance of speaker verification from the original models without adjusting the structure and hyper-parameters of them, suggesting that adversarially disentangled representation is extremely useful for alleviating the lack of speaker labels.
Learning a Facial Expression Embedding Disentangled from Identity
TLDR
This paper proposes a Deviation Learning Network (DLN) with a pseudo-siamese structure to extract the deviation feature vector, and provides high-order polynomial to nonlinearly project the high-dimensional feature to a low-dimensional manifold.
Decorrelated Adversarial Learning for Age-Invariant Face Recognition
TLDR
A novel algorithm to remove age-related components from features mixed with both identity and age information is presented, which learns the decomposed features of age and identity whose correlation is significantly reduced.
Adversarial Learning for Age-Invariant Face Recognition
TLDR
A novel algorithm to remove age-related components from features mixed with both identity and age information by factorizing a mixed face feature into two uncorrelated components: identity-dependent component and age- dependent component, where the identitydependent component contains information that is useful for face recognition.
...
...

References

SHOWING 1-10 OF 52 REFERENCES
Disentangled Representation Learning GAN for Pose-Invariant Face Recognition
TLDR
Quantitative and qualitative evaluation on both controlled and in-the-wild databases demonstrate the superiority of DR-GAN over the state of the art.
Reconstruction for Feature Disentanglement in Pose-invariant Face Recognition
TLDR
This paper presents a method for learning a feature representation that is invariant to pose, without requiring extensive pose coverage in training data, and proposes a Siamese network to explicitly disentangle identity and pose.
Representation Learning by Rotating Your Faces
TLDR
A Disentangled Representation learning-Generative Adversarial Network (DR-GAN) with three distinct novelties that demonstrate the superiority of DR-GAN over the state of the art in both learning representations and rotating large-pose face images.
Deep Learning Face Representation by Joint Identification-Verification
TLDR
This paper shows that the face identification-verification task can be well solved with deep learning and using both face identification and verification signals as supervision, and the error rate has been significantly reduced.
Conditional CycleGAN for Attribute Guided Face Image Generation
TLDR
This work extends the cycleGAN to Conditional cycleGAN such that the mapping from X to Y is subjected to attribute condition Z, and uses face feature vector extracted from face verification network as Z to demonstrate the efficacy of this approach on identity preserving face image super-resolution.
Surpassing Human-Level Face Verification Performance on LFW with GaussianFace
TLDR
A principled multi-task learning approach based on Discriminative Gaussian Process Latent VariableModel (DGPLVM), named GaussianFace, for face verification, which achieved an impressive accuracy rate and introduced a more efficient equivalent form of Kernel Fisher Discriminant Analysis to DGPLVM.
Towards Large-Pose Face Frontalization in the Wild
TLDR
This work proposes a novel deep 3D Morphable Model (3DMM) conditioned Face Frontalization Generative Adversarial Network (GAN), termed as FF-GAN, to generate neutral head pose face images, which differs from both traditional GANs and 3DMM based modeling.
Deep Identity-aware Transfer of Facial Attributes
TLDR
A Deep convolutional network model for Identity-Aware Transfer of facial attributes that can provide a unified solution for several representative facial attribute transfer tasks, e.g., expression transfer, accessory removal, age progression, and gender transfer, and can be extended for other face enhancement tasks such as face hallucination.
Deep Learning Face Attributes in the Wild
TLDR
A novel deep learning framework for attribute prediction in the wild that cascades two CNNs, LNet and ANet, which are fine-tuned jointly with attribute tags, but pre-trained differently.
Beyond Face Rotation: Global and Local Perception GAN for Photorealistic and Identity Preserving Frontal View Synthesis
TLDR
A Two-Pathway Generative Adversarial Network (TP-GAN) for photorealistic frontal view synthesis by simultaneously perceiving global structures and local details and outperforms state-of-theart results on large pose face recognition.
...
...