DPMambaIR: All-in-One Image Restoration via Degradation-Aware Prompt State Space Model
Zhanwen Liu, Sai Zhou, Yuchao Dai, Yang Wang † , Yisheng An, Xiangmo Zhao · 2025
Read DPMambaIR: All-in-One Image Restoration via Degradation-Aware Prompt State Space Model by Zhanwen Liu, Sai Zhou, Yuchao Dai, Yang Wang † , Yisheng An, Xiangmo Zhao (2025) 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/arxiv-2504-17732.txt.
Description
Zhanwen Liu and co-authors present a 2025 model that fixes many kinds of image damage with one system. Photos can suffer blur, noise, rain, or low light, and past tools often needed one model per flaw. Prompt-guided tools exist, but their prompts stay coarse. The team's model reads fine detail about the damage and lets that detail steer the repair.
A damage reader pulls fine traits from each image. Those traits become live prompts inside a state space model that rebuilds the image. A second block restores fine local detail that broad repair can lose. On a mixed set with seven damage kinds, the model scored best on two standard quality scales. One tool for all damage types suits phone and archive repair better than a shelf of single fixes.
Catalog record
| CREATOR | Zhanwen Liu, Sai Zhou, Yuchao Dai, Yang Wang † , Yisheng An, Xiangmo Zhao |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 53 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-17732