Adam Kowalski
Copyright © 2018 Adam Kowalski
This project will try to look at medical images and determine whether or not they contain melanoma. Through the use of deep learning, I will build models that try to extract features from the images so that I can correctly classify the images.
Reproducability of results is a huge concern in the field of deep learning, so I will use docker to ensure environmental differences between developers are kept to a minimum. There will be containers for getting the dataset, as well as building, training and evaluating models.
Another cruicial aspect of developing models is tuning hyper parameters. It is important to change them in a controlled way, so that you can understand which models performed better and get an intution into how to change them in the future so that they may keep getting better.
https://github.com/adam-r-kowalski/melanoma-classifier/issues
git clone https://github.com/adam-r-kowalski/melanoma-classifier.git
cd melanoma-classifier
In order to ensure no environmental differences between developers, docker is used. Follow the installation instructions for your platform.
Because we want to take advantage of the GPU to make training much faster, we need to also install nvidia-docker.
This service will download the dataset from The International Skin Imaging Collaboration who generously provide thousands of examples of various skin related diseases.
docker-compose -f data-downloader/docker-compose.yml up
This service will take our data and create an efficient binary representation which contains both the images as well as the labels.
docker-compose -f data-labeler/docker-compose.yml up
This service will take our dataset and partition it into batches of 1000 as well as ensure that we have an even distribution of melanoma to non melanoma images
docker-compose -f data-partitioner/docker-compose.yml up
This service will allow you to construct and train models
docker-compose -f webapp/docker-compose.yml up
This code is available under the "Apache License 2.0". Please see the file COPYING in this distribution for license terms.