Ensuring data protection for eye images by combining fine-tuned image synthesis and anonymity assessment

Sebastian Uschmann, Cord Spreckelsen, Sven Festag · 2026

Read Ensuring data protection for eye images by combining fine-tuned image synthesis and anonymity assessment by Sebastian Uschmann, Cord Spreckelsen, Sven Festag (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/PMC13395446.txt.

Description

Sebastian Uschmann, Cord Spreckelsen, and Sven Festag build, in 2026, a guard for sharing fake eye photos safely. AI image makers can craft real-looking eyes for teaching and research, yet fakes may leak the identity of real training faces. Their two-stage system first makes eye images with a tuned Stable Diffusion model. A privacy scorer then flags fakes that sit too close to real people.

Tests used two thousand photos from 704 people. The scorer caught more than ninety-seven percent of risky fakes. Safe fakes looked less real than risky ones, which shows the trade-off plainly. The authors pitch the frame for clinics that must share images without naming patients.

Catalog record

CREATORSebastian Uschmann, Cord Spreckelsen, Sven Festag
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE39889
LICENSEhttps://creativecommons.org/licenses/by/4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/article/MED/PMC13395446

Browse the full library · About the Tower of Babel