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

TL;DR

DNA-GAN introduces a supervised model that learns disentangled, identity-preserving image representations by encoding attributes as DNA-like segments, enabling attribute manipulation and realistic image generation.

Contribution

The paper presents a novel DNA-GAN model that effectively disentangles image attributes using a DNA-like encoding, improving over existing methods in quality and controllability.

Findings

Effective disentangling of multiple image attributes.

Generation of realistic images with attribute modifications.

Outperforms some existing methods on Multi-PIE and CelebA datasets.

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.