Please use this identifier to cite or link to this item:
Title: EAD-GAN: a generative adversarial network for disentangling affine transforms in images
Authors: Liu, Letao
Jiang, Xudong
Saerbeck, Martin
Dauwels, Justin
Keywords: Engineering::Electrical and electronic engineering
Issue Date: 2022
Source: Liu, L., Jiang, X., Saerbeck, M. & Dauwels, J. (2022). EAD-GAN: a generative adversarial network for disentangling affine transforms in images. IEEE Transactions On Neural Networks and Learning Systems, PP, 1-11.
Project: S17-1298-IPP-II
Journal: IEEE transactions on neural networks and learning systems
Abstract: This article proposes a generative adversarial network called explicit affine disentangled generative adversarial network (EAD-GAN), which explicitly disentangles affine transform in a self-supervised manner. We propose an affine transform regularizer to force the InfoGAN to have explicit properties of affine transform. To facilitate training an affine transform encoder, we decompose the affine matrix into two separate matrices and infer the explicit transform parameters by the least-squares method. Unlike the existing approaches, representations learned by the proposed EAD-GAN have clear physical meaning, where transforms, such as rotation, horizontal and vertical zooms, skews, and translations, are explicitly learned from training data. Thus, we set different values of each transform parameter individually to generate specifically affine transformed data by the learned network. We show that the proposed EAD-GAN successfully disentangles these attributes on the MNIST, CelebA, and dSprites datasets. EAD-GAN achieves higher disentanglement scores with a large margin compared to the state-of-the-art methods on the dSprites dataset. For example, on the dSprites dataset, EAD-GAN achieves the MIG and DCI score of 0.59 and 0.96 respectively, compared to 0.37 and 0.71, respectively, for the state-of-the-art methods.
ISSN: 2162-237X
DOI: 10.1109/TNNLS.2022.3195533
Schools: School of Electrical and Electronic Engineering 
Rights: © 2022 The authors. This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see
Fulltext Permission: open
Fulltext Availability: With Fulltext
Appears in Collections:EEE Journal Articles

Files in This Item:
File Description SizeFormat 
EAD-GAN_A_Generative_Adversarial_Network_for_Disentangling_Affine_Transforms_in_Images.pdf4.57 MBAdobe PDFThumbnail

Citations 50

Updated on Apr 8, 2024

Page view(s)

Updated on Apr 14, 2024


Updated on Apr 14, 2024

Google ScholarTM




Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.