RobMedNAS: searching robust neural network architectures for medical image synthesis.
| Autor: | Zhang J; Electrical and Computer Engineering, University of Wisconsin-Madison, Madison, United States of America., Chen W; Electrical and Computer Engineering, University of Wisconsin-Madison, Madison, United States of America., Joshi T; Electrical and Computer Engineering, University of Wisconsin-Madison, Madison, United States of America., Uyanik M; Medical Physics, University of Wisconsin-Madison, Madison, United States of America., Zhang X; Computer Science, University of Wisconsin-Madison, Madison, United States of America., Loh PL; Pure Mathematics and Mathematical Statistics, University of Cambridge, Cambridge, United Kingdom., Jog V; Pure Mathematics and Mathematical Statistics, University of Cambridge, Cambridge, United Kingdom., Bruce R; Radiology, University of Wisconsin-Madison, Madison, United States of America., Garrett J; Radiology, University of Wisconsin-Madison, Madison, United States of America., McMillan A; Electrical and Computer Engineering, University of Wisconsin-Madison, Madison, United States of America.; Medical Physics, University of Wisconsin-Madison, Madison, United States of America.; Radiology, University of Wisconsin-Madison, Madison, United States of America.; Biomedical Engineering, University of Wisconsin-Madison, Madison, United States of America. |
|---|---|
| Jazyk: | angličtina |
| Zdroj: | Biomedical physics & engineering express [Biomed Phys Eng Express] 2024 Aug 23; Vol. 10 (5). Date of Electronic Publication: 2024 Aug 23. |
| DOI: | 10.1088/2057-1976/ad6e87 |
| Abstrakt: | Investigating U-Net model robustness in medical image synthesis against adversarial perturbations, this study introduces RobMedNAS, a neural architecture search strategy for identifying resilient U-Net configurations. Through retrospective analysis of synthesized CT from MRI data, employing Dice coefficient and mean absolute error metrics across critical anatomical areas, the study evaluates traditional U-Net models and RobMedNAS-optimized models under adversarial attacks. Findings demonstrate RobMedNAS's efficacy in enhancing U-Net resilience without compromising on accuracy, proposing a novel pathway for robust medical image processing. (Creative Commons Attribution license.) |
| Databáze: | MEDLINE |
| Externí odkaz: |
|