Unsupervised content and style learning for multimodal cross-domain image translation
Zhijie Lin, Jingjing Chen, Xiaolong Ma, Chao Li, Huiming Zhang, Lei Zhao · 2024
Read Unsupervised content and style learning for multimodal cross-domain image translation by Zhijie Lin, Jingjing Chen, Xiaolong Ma, Chao Li, Huiming Zhang, Lei Zhao (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-11603042.txt.
Description
Lin and co-authors present a way to restyle an image while keeping its shapes intact, without needing paired training examples. Image translation can turn, say, a photo into a painting-style version, but many tools distort colors or warp the content while doing so.
Their method learns content and style apart. A self-structure attention loss guards the layout of the scene, while a color constraint loss pulls the colors toward those of a reference image. Tests on several datasets beat earlier methods on standard quality scores and kept both structure and color patterns better. Clean separation of shape and style is useful for design, film, and photo tools.
Catalog record
| CREATOR | Zhijie Lin, Jingjing Chen, Xiaolong Ma, Chao Li, Huiming Zhang, Lei Zhao |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 82 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |