Rethinking Privacy in Medical Imaging AI: From Metadata and Pixel-Level Identification Risks to Federated Learning and Synthetic Data Challenges
Konstantina Giouroukou, Kostas Marias, Manolis Tsiknakis, Michail E. Klontzas · Radiology Artificial Intelligence · 2025
This report reviews methods for preparing imaging data for artificial intelligence applications, focusing on the need for robust privacy protection through de-identification, federated learning, and synthetic data generation, while highlighting the potential risks associated with these approaches.