Analytical Logit Scaling for High-Resolution Sea Ice Topology Retrieval from Weakly Labeled SAR Imagery
Reda Elwaradi, Julien Gimenez, Stéphane Hordoir, and Mehdi Ait Hamma, Adrien Chan Hon Tong * and Flora Weissgerber * · 2026
Read Analytical Logit Scaling for High-Resolution Sea Ice Topology Retrieval from Weakly Labeled SAR Imagery by Reda Elwaradi, Julien Gimenez, Stéphane Hordoir, and Mehdi Ait Hamma, Adrien Chan Hon Tong * and Flora Weissgerber * (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/arxiv-2603-13573.txt.
Description
Reda Elwaradi and co-authors present a weakly supervised pipeline for high-resolution sea ice topology retrieval from SAR imagery. Operational ice charts give only coarse region-level polygons, so models struggle for pixel-level accuracy. Their U-Net fuses Sentinel-1 SAR and AMSR-2 radiometry data with a region-based loss. An Analytical Logit Scaling step then fixes under-confident outputs after inference. It sets temperature and bias from latent space percentiles of each scene. The method reveals fine sea ice fractures at 40-meter resolution with no manual pixel labels.
Catalog record
| CREATOR | Reda Elwaradi, Julien Gimenez, Stéphane Hordoir, and Mehdi Ait Hamma, Adrien Chan Hon Tong * and Flora Weissgerber * |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 18762 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.13573