Interpretable machine learning via symbolic classification of radiomic texture and morphological features for pediatric pneumonia detection from chest X-rays
Filippos Sofos, Zoi D Pana, Dimitris Drikakis · 2026
Read Interpretable machine learning via symbolic classification of radiomic texture and morphological features for pediatric pneumonia detection from chest X-rays by Filippos Sofos, Zoi D Pana, Dimitris Drikakis (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/plos-pone-0351081.txt.
Description
Filippos Sofos, Zoi Pana, and Dimitris Drikakis built a pneumonia detector for children's chest X-rays that shows its reasoning as one readable formula. Most image AI hides its logic inside huge networks, leaving doctors with a verdict they cannot inspect. Their approach evolves a compact equation from a handful of texture and shape measures taken from the X-ray, measures a radiologist can recognize, like entropy and solidity.
The formula sorted the main dataset with accuracy in the high eighties and held up on a filtered independent test set at just under eighty percent. It was also checked and retuned on a dataset from another source. With vastly fewer numbers inside than a deep network, the equation can be printed, debated, and audited, a different bargain for clinics weighing trust against raw accuracy.
Catalog record
| CREATOR | Filippos Sofos, Zoi D Pana, Dimitris Drikakis |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 52799 |
| LICENSE | CC_BY |
| RIGHTS | CC BY |
| SOURCE | PLOS |