Skip to content
This repository has been archived by the owner on May 6, 2023. It is now read-only.

gkaissis/PriMIA

Repository files navigation

PriMIA

Privacy-preserving Medical Image Analysis

About

PriMIA is a framework for end-to-end privacy preserving machine learning on medical images designed and built at the Technical University of Munich, OpenMined and Imperial College London. It enables you to do federated training of convolutional neural networks and encrypted inference from a simple command-line interface.

Documentation

Go here!

Dataset

Please check this file for information on how to use this dataset.


The dataset in this repository is being re-used under the license terms from the CoronaHack Chest X-Ray Dataset on Kaggle and modified as indicated in Dataset_Description.md.

Original citation for the majority of the images: https://data.mendeley.com/datasets/rscbjbr9sj/2

CC BY 4.0

This work is licensed under a Creative Commons Attribution 4.0 International License.

CC BY 4.0

PySyft, included with PriMIA is licensed under the Apache 2.0 license. The license can be found here. All modifications to the PySyft source code are also licensed under the Apache 2.0 license.

The PriMIA source code is licensed under the Attribution-NonCommercial-ShareAlike 4.0 International License. (C) 2020, PriMIA contributors.

About

PriMIA: Privacy-preserving Medical Image Analysis

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages