Solving Inverse Problems with FLAIR
Julius Erbach, Dominik Narnhofer, Andreas Dombos, Bernt Schiele, Jan Eric Lenssen, Konrad Schindler, [0.5em] ETH Zürich · 2025
Read Solving Inverse Problems with FLAIR by Julius Erbach, Dominik Narnhofer, Andreas Dombos, Bernt Schiele, Jan Eric Lenssen, Konrad Schindler, [0.5em] ETH Zürich (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-2506-02680.txt.
Description
Julius Erbach and co-authors wrote a 2025 paper on FLAIR, a method that uses flow based image generators as priors to undo damage in images. Rebuilding a clean image from a blurred or partial one is an inverse problem. Past generator based fixes drift from the measured data or clash with it.
FLAIR splits the job: one term enforces exact match to the data, another steers toward likely images. A time varying dial sets how hard the prior pulls. No retraining is needed per damage type. On standard tests it beat diffusion and flow rivals in quality and variety alike.
Catalog record
| CREATOR | Julius Erbach, Dominik Narnhofer, Andreas Dombos, Bernt Schiele, Jan Eric Lenssen, Konrad Schindler, [0.5em] ETH Zürich |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 52 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-02680