STACOM 2018 Challenge Code

This is the code used for the submission to the LVQuant challenge and published as Left-Ventricle Quantification Using Residual U-Net. It relies on PyTorch 0.4.0+, Numpy, Scipy, and the included submodule DeepLearnUtils. When checking out this repo you may have to run the following to get the submodule:

git submodule update --init --recursive

To start training, first copy the train data cardiac-dig.mat and test data lvquan_test_images_30sub.mat files into this directory. Run deformclone.py to generate the augmented training dataset. Training of the model is then done in STACOMTrain.ipynb and evaluation of metrics in STACOMTestEval.ipynb.