Is synthetic data generation effective in maintaining clinical biomarkers? Investigating diffusion models across diverse imaging modalities
Abdullah Hosseini, Ahmed Serag · 2025
Read Is synthetic data generation effective in maintaining clinical biomarkers? Investigating diffusion models across diverse imaging modalities by Abdullah Hosseini, Ahmed Serag (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/pmc-11826350.txt.
Description
Abdullah Hosseini and Ahmed Serag published this study, dated 2024 in the head, on whether fake medical images keep the clues doctors rely on. Privacy rules block sharing real scans, so labs make synthetic ones with AI. A diffusion model, which builds images by clearing noise, did the making here. The open question was whether hidden disease markers survive the faking.
The team used a transformer-based network across three image types: X-ray style radiology, eye scans, and tissue slides. Fake images kept features like lung markings and retinal flaws. A classifier trained to spot markers scored high on the fakes, with key measures between 0.8 and 0.99. The results suggest synthetic sets can support research without exposing patients.
Catalog record
| CREATOR | Abdullah Hosseini, Ahmed Serag |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 25 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |