This is the code for an undegraduate research program at DePaul University.
It contains code for generating 3D lung nodule ct scans From LIDC data through a Wasserstein Generative Adversarial Network with Gradient Penalty (WGAN-GP).
The data can be found at https://wiki.cancerimagingarchive.net/display/Public/LIDC-IDRI (this is large, ~100 gb).
Some data preprocessing files are not present in this repository, so you may have to write your own (getting and sorting the dicom images into a dictionary).
The paper is published at SPIE Medical Imaging: Augmenting LIDC Dataset Using 3D Generative Adversarial Networks to Improve Lung Nodule Detection (pdf available on ResearchGate)
If you do decide to use this, please cite:
@inproceedings{gao2019augmenting,
title={Augmenting LIDC dataset using 3D generative adversarial networks to improve lung nodule detection},
author={Gao, Chufan and Clark, Stephen and Furst, Jacob and Raicu, Daniela},
booktitle={Medical Imaging 2019: Computer-Aided Diagnosis},
volume={10950},
pages={109501K},
year={2019},
organization={International Society for Optics and Photonics}
}