StyleGAN-∞: Extending StyleGAN to arbitrary-ratio translation with StyleBook

Publication Type

Journal Article

Publication Date

9-2025

Abstract

Although pre-trained large-scale generative models StyleGAN series have proven to be effective in various editing and translation tasks, they are limited to pre-defined fixed aspect ratio. To overcome this limitation, we propose StyleGAN-∞, a model that enables pre-trained StyleGAN to perform arbitrary-ratio conditional synthesis. Our key insight is to distill the expressive StyleGAN features into a StyleBook, such that an arbitrary-ratio condition can be translated to other forms by properly assembling pre-defined StyleBook vectors. To learn and leverage the StyleBook, we employ a network with three distinct stages, each corresponding to StyleBook extraction, StyleBook correspondence learning, and arbitrary-ratio synthesis. Extensive experiments on various conditional synthesis tasks, like super-resolution, sketch synthesis, and semantic synthesis, demonstrate superior performances over state-of-the-art image-to-image translation methods. Moreover, our model can easily generate megapixel images in diverse modalities by taking advantage of different pre-trained StyleGAN models.

Keywords

Generative adversarial networks, image-to-image translation, conditional synthesis

Discipline

Graphics and Human Computer Interfaces | Software Engineering

Research Areas

Software and Cyber-Physical Systems

Publication

IEEE Transactions on Visualization and Computer Graphics

Volume

31

Issue

9

First Page

6575

Last Page

6587

ISSN

1077-2626

Identifier

10.1109/TVCG.2024.3522565

Publisher

Institute of Electrical and Electronics Engineers

Additional URL

https://doi.org/10.1109/TVCG.2024.3522565

This document is currently not available here.

Share

COinS