This is supplementary material for the manuscript:

"Lung Pattern Classification for Interstitial Lung Diseases Using a Deep Convolutional Neural Network"
M. Anthimopoulos, S. Christodoulidis, L. Ebner, A. Christe and S. Mougiakakou
IEEE Transactions on Medical Imaging (2016)
http://dx.doi.org/10.1109/TMI.2016.2535865

In case of any questions, please do not hesitate to contact us.

This code was used on a Linux machine with Ubuntu (14.04.3 LTS) using the following setup:

• CUDA (7.5)
• cuDNN (v3)
• python (2.7.6) with
  • Theano (0.8)
  • keras (0.3.2)
  • numpy (1.10.4)
  • argparse (1.2.1)
  • scikit-learn (0.17)
• python-opencv (2.4.10)

There are three major components

• main.py : the main script which parses the train parameters, loads some sample data and runs the training of the CNN.
• helpers.py : a file with some helper functions for parsing the input parameters, loading sample data and calculating a number of evaluation metrics
• cnn_model.py : this file implements the architecture of the proposed CNN and trains it.

python main.py : runs an experiment with the default parameters
python main.py -h : shows the help message

The execution outputs two csv formatted files with the performance metrics of the CNN. The first contains the performances for each training epoch while the second only for the epochs that improved the performance. The code prints the same output while running as well as a confusion matrix every time the CNN performance improves.

Copyright (C) 2016 Marios Anthimopoulos, Stergios Christodoulidis, Stavroula Mougiakakou / University of Bern

This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program. If not, see http://www.gnu.org/licenses/.