Imperial College London
Alternate

ProfessorDanielRueckert

Faculty of Engineering, Department of Computing

Professor of Visual Information Processing
 
 
 
//

Contact

 

+44 (0)20 7594 8333d.rueckert Website

 
 
//

Location

 

568Huxley BuildingSouth Kensington Campus

//

Summary

 

Publications

Citation

BibTex format

@article{Kaissis:2021:10.1038/s42256-021-00337-8,
author = {Kaissis, G and Ziller, A and Passerat-Palmbach, J and Ryffel, T and Usynin, D and Trask, A and Lima, I and Mancuso, J and Jungmann, F and Steinborn, M-M and Saleh, A and Makowski, M and Rueckert, D and Braren, R},
doi = {10.1038/s42256-021-00337-8},
journal = {Nature Machine Intelligence},
pages = {473--484},
title = {End-to-end privacy preserving deep learning on multi-institutional medical imaging},
url = {http://dx.doi.org/10.1038/s42256-021-00337-8},
volume = {3},
year = {2021}
}
Download

RIS format (EndNote, RefMan)

TY  - JOUR
AB  - Using large, multi-national datasets for high-performance medical imaging AI systems requires innovation in privacy-preserving machine learning so models can train on sensitive data without requiring data transfer. Here we present PriMIA (Privacy-preserving Medical Image Analysis), a free, open-source software framework for differentially private, securely aggregated federated learning and encrypted inference on medical imaging data. We test PriMIA using a real-life case study in which an expert-level deep convolutional neural network classifies paediatric chest X-rays; the resulting model’s classification performance is on par with locally, non-securely trained models. We theoretically and empirically evaluate our framework’s performance and privacy guarantees, and demonstrate that the protections provided prevent the reconstruction of usable data by a gradient-based model inversion attack. Finally, we successfully employ the trained model in an end-to-end encrypted remote inference scenario using secure multi-party computation to prevent the disclosure of the data and the model.
AU  - Kaissis,G
AU  - Ziller,A
AU  - Passerat-Palmbach,J
AU  - Ryffel,T
AU  - Usynin,D
AU  - Trask,A
AU  - Lima,I
AU  - Mancuso,J
AU  - Jungmann,F
AU  - Steinborn,M-M
AU  - Saleh,A
AU  - Makowski,M
AU  - Rueckert,D
AU  - Braren,R
DO  - 10.1038/s42256-021-00337-8
EP  - 484
PY  - 2021///
SN  - 2522-5839
SP  - 473
TI  - End-to-end privacy preserving deep learning on multi-institutional medical imaging
T2  - Nature Machine Intelligence
UR  - http://dx.doi.org/10.1038/s42256-021-00337-8
UR  - https://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=PARTNER_APP&SrcAuth=LinksAMR&KeyUT=WOS:000655137800001&DestLinkType=FullRecord&DestApp=ALL_WOS&UsrCustomerID=a2bf6146997ec60c407a63945d4e92bb
UR  - https://www.nature.com/articles/s42256-021-00337-8
VL  - 3
ER  -
Download