Phys-Mamba: Physics-informed selective state-space fusion network for high-fidelity underwater image restoration
Hongwei Li, Weishen Li, Yuqing Lao, Jie Li, Junjie He, Caihong Wu, Huilong Zhong, Shaoji Huang · 2026
Read Phys-Mamba: Physics-informed selective state-space fusion network for high-fidelity underwater image restoration by Hongwei Li, Weishen Li, Yuqing Lao, Jie Li, Junjie He, Caihong Wu, Huilong Zhong, Shaoji Huang (2026) 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/plos-pone-0354030.txt.
Description
Hongwei Li and co-authors build an AI system that cleans up underwater photographs blurred and tinted by water itself. Robots and divers need clear images, but many enhancers either invent detail or run too slowly on large frames. Their design steers a fast sequence model with variables drawn from the physics of underwater light loss, and merges detail across scales without letting the model balloon in size.
Across four public test sets, the system scored near the top on standard quality measures while staying compact. The authors report it handled even very large frames thanks to its step-by-step cost growing gently with image size. Li and colleagues pitch the design for subsea robots that must see clearly in real time.
Catalog record
| CREATOR | Hongwei Li, Weishen Li, Yuqing Lao, Jie Li, Junjie He, Caihong Wu, Huilong Zhong, Shaoji Huang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 43432 |
| LICENSE | CC_BY |
| RIGHTS | CC BY |
| SOURCE | PLOS |