Adaptive Optics Image Analysis Using Generative AI (GPT-4) Scripting: Exploratory Study
Jeroen AAH Pas, Ludo C van der Zanden, Yoeri van Leeuwen, Yara TE Lechanteur, Carel B Hoyng, Thomas Theelen · 2026
Read Adaptive Optics Image Analysis Using Generative AI (GPT-4) Scripting: Exploratory Study by Jeroen AAH Pas, Ludo C van der Zanden, Yoeri van Leeuwen, Yara TE Lechanteur, Carel B Hoyng, Thomas Theelen (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/pmc-13620430.txt.
Description
Adaptive optics ophthalmoscopy images the retina's cone cells, but analysis scripts for it are scarce. That gap bites hardest for diseased eyes and large batches. The six authors ask whether a large language model can write the analysis code. Using GPT-4, they iteratively built an R script for flood illumination images. They refined it through 54 rounds of instructions against images from healthy and Stargardt disease participants.
A researcher blind to the process then checked the script on a fresh set of images. Its cone counts were compared with two human graders and with existing software trained on healthy eyes. The script worked: it found and counted the blobs that mark cones. The authors stress it is a nonvalidated proof of principle. Much more tuning and testing is needed before such scripts can supply trial biomarkers.
Catalog record
| CREATOR | Jeroen AAH Pas, Ludo C van der Zanden, Yoeri van Leeuwen, Yara TE Lechanteur, Carel B Hoyng, Thomas Theelen |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 29 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |