Author: Paras Lakhani, paras.lakhani@jefferson.edu
More details and a step-by-step guide for the tutorial can be found in the Journal of Digital Imaging Publication (ref: ), which is the official journal of the Society of Imaging Informatics in Medicine (SIIM).
This is a high-level introduction into practical machine learning for purposes of medical image classification.
In this tutorial, we use the Tensorflow framework as it is currently the most actively used and the Keras library, which a high-level application programming interface that simplifies working with Tensorflow.
We hope that this tutorial will spark interest and provide a basic starting point for those interested in machine learning in regard to medical imaging.
A Jupyter ipython notebook is provided called "HelloWorldDeepLearning.ipynb"
We provide 75 images, 38 are chest X-rays, and 37 are abdominal X-rays. These de-identified PNGs obtained from openI, https://openi.nlm.nih.gov/, a searchable online repository of medical images from published PubMed Central articles
The goal of this tutorial is to build a deep learning classifier to accurately differentiate between the two.
You'll need a computer with the following installed:
- Tensorflow (https://www.tensorflow.org)
- Keras library (https://keras.io)
- Jupyter (http://jupyter.org)
- Download the x-rays provided in .zip file
To make things easier, there is a convenient SIIM docker that has Tensorflow / Keras / Jupyterlab already installed, located here: https://github.com/ImagingInformatics/machine-learning/tree/master/docker-keras-tensorflow-python3-jupyter
After your environment is set up, open the ipython notebook, and run the code!