Enabling PSO-Secure Synthetic Data Sharing Using Diversity-Aware Diffusion Models
Dombrowski MN, Kainz B (2025)
Publication Language: English
Publication Type: Conference contribution
Publication year: 2025
Journal
Publisher: Springer
Series: Lecture Notes in Computer Science
City/Town: Cham
Book Volume: 16135
Pages Range: 25-35
Conference Proceedings Title: Bridging Regulatory Science and Medical Imaging Evaluation; and Distributed, Collaborative, and Federated Learning. First International Workshop, BRIDGE 2025, and 6th International Workshop, DeCaF 2025, Held in Conjunction with MICCAI 2025, Daejeon, South Korea, September 23 and September 27, 2025, Proceedings
Event location: Daejeon
ISBN: 9783032056658
DOI: 10.1007/978-3-032-05663-4_3
Abstract
Synthetic data has recently reached a level of visual fidelity that makes it nearly indistinguishable from real data, offering great promise for privacy-preserving data sharing in medical imaging. However, fully synthetic datasets still suffer from significant limitations: First and foremost, the legal aspect of sharing synthetic data is often neglected and data regulations, such as the GDPR, are largley ignored. Secondly, synthetic models fall short of matching the performance of real data, even for in-domain downstream applications. Recent methods for image generation have focused on maximising image diversity instead of fidelity solely to improve the mode coverage and therefore the downstream performance of synthetic data. In this work, we shift perspective and highlight how maximizing diversity can also be interpreted as protecting natural persons from being singled out, which leads to predicate singling-out (PSO) secure synthetic datasets. Specifically, we propose a generalisable framework for training diffusion models on personal data which leads to unpersonal synthetic datasets achieving performance within one percentage point of real-data models while significantly outperforming state-of-the-art methods that do not ensure privacy. Our code is available at https://github.com/MischaD/Trichotomy.
Authors with CRIS profile
Mischa Neil Dombrowski
Professur für Image Data Exploration and Analysis
Bernhard Kainz
Professur für Image Data Exploration and Analysis
How to cite
APA:
Dombrowski, M.N., & Kainz, B. (2025). Enabling PSO-Secure Synthetic Data Sharing Using Diversity-Aware Diffusion Models. In Ghada Zamzmi, Annika Reinke, Ravi Samala, Meirui Jiang, Xiaoxiao Li, Holger Roth, Mariia Sidulova, Thijs Kooi, Shadi Albarqouni, Spyridon Bakas, Nicola Rieke (Eds.), Bridging Regulatory Science and Medical Imaging Evaluation; and Distributed, Collaborative, and Federated Learning. First International Workshop, BRIDGE 2025, and 6th International Workshop, DeCaF 2025, Held in Conjunction with MICCAI 2025, Daejeon, South Korea, September 23 and September 27, 2025, Proceedings (pp. 25-35). Daejeon, KR: Cham: Springer.
MLA:
Dombrowski, Mischa Neil, and Bernhard Kainz. "Enabling PSO-Secure Synthetic Data Sharing Using Diversity-Aware Diffusion Models." Proceedings of the 1st International Workshop on Bridging Regulatory Science and Medical Imaging Evaluation, BRIDGE 2025 and 6th MICCAI Workshop on Distributed, Collaborative and Federated Learning, DeCaF 2025, Held in Conjunction with 28th International conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2025, Daejeon Ed. Ghada Zamzmi, Annika Reinke, Ravi Samala, Meirui Jiang, Xiaoxiao Li, Holger Roth, Mariia Sidulova, Thijs Kooi, Shadi Albarqouni, Spyridon Bakas, Nicola Rieke, Cham: Springer, 2025. 25-35.
BibTeX: Download