DNA-GAN: Learning Disentangled Representations from Multi-Attribute Images

Abstract

Disentangling factors of variation has become a very challenging problem on representation learning. Existing algorithms suffer from many limitations, such as unpredictable disentangling factors, poor quality of generated images from encodings, lack of identity information, etc. In this paper, we propose a supervised learning model called DNA-GAN which tries to disentangle different factors or attributes of images. The latent representations of images are DNA-like, in which each individual piece (of the encoding) represents an independent factor of the variation. By annihilating the recessive piece and swapping a certain piece of one latent representation with that of the other one, we obtain two different representations which could be decoded into two kinds of images with the existence of the corresponding attribute being changed. In order to obtain realistic images and also disentangled representations, we further introduce the discriminator for adversarial training. Experiments on Multi-PIE and CelebA datasets finally demonstrate that our proposed method is effective for factors disentangling and even overcome certain limitations of the existing methods.

ICLR 2018 workshop, github: https://github.com/Prinsphield/DNA-GAN

References in corpus (3)

• Deep Convolutional Inverse Graphics Network
• Invertible Conditional GANs for image editing
• Unsupervised Cross-Domain Image Generation

Cited by in corpus (11)

• ELEGANT: Exchanging Latent Encodings with GAN for Transferring Multiple Face Attributes
• Semi-Supervised StyleGAN for Disentanglement Learning
• Dual Swap Disentangling
• Disentangling Multiple Conditional Inputs in GANs
• ChildPredictor: A Child Face Prediction Framework with Disentangled Learning
• Disentangling Factors of Variation with Cycle-Consistent Variational Auto-Encoders
• Capturing Label Characteristics in VAEs
• A Survey of Deep Facial Attribute Analysis
• Covering up bias in CelebA-like datasets with Markov blankets: A post-hoc cure for attribute prior avoidance
• ClsGAN: Selective Attribute Editing Model Based On Classification Adversarial Network
• BSD-GAN: Branched Generative Adversarial Network for Scale-Disentangled Representation Learning and Image Synthesis