DSA-CycleGAN: A Domain Shift Aware CycleGAN for Robust Multi-Stain Glomeruli Segmentation
Zeeshan Nisar, Friedrich Feuerhake, Thomas Lampert · 2026
Read DSA-CycleGAN: A Domain Shift Aware CycleGAN for Robust Multi-Stain Glomeruli Segmentation by Zeeshan Nisar, Friedrich Feuerhake, Thomas Lampert (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-2604-18368.txt.
Description
Nisar, Feuerhake, and Lampert propose DSA-CycleGAN, a stain-transfer method for segmenting glomeruli in digital pathology slides. Stain variation degrades models, and labeling every stain costs too much time and money. Standard CycleGAN adds noise when one stain maps to many, clashing with its cycle-consistency loss. The domain-shift-aware version cuts that noise and beats rivals, most clearly when translating between biologically distinct stains.
Catalog record
| CREATOR | Zeeshan Nisar, Friedrich Feuerhake, Thomas Lampert |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 19614 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.18368