Dai, Yihua, Xiang, Tianyi, Deng, Bailin ![]() ![]() |
Preview |
PDF
- Accepted Post-Print Version
Download (60MB) | Preview |
Preview |
PDF
- Supplemental Material
Download (58MB) | Preview |
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.
Item Type: | Article |
---|---|
Date Type: | Published Online |
Status: | In Press |
Schools: | Computer Science & Informatics |
Subjects: | Q Science > QA Mathematics > QA75 Electronic computers. Computer science Q Science > QA Mathematics > QA76 Computer software |
Publisher: | Institute of Electrical and Electronics Engineers |
ISSN: | 1077-2626 |
Date of First Compliant Deposit: | 24 December 2024 |
Date of Acceptance: | 11 December 2024 |
Last Modified: | 16 Jan 2025 16:30 |
URI: | https://orca.cardiff.ac.uk/id/eprint/174909 |
Actions (repository staff only)
![]() |
Edit Item |