This repository is based on the cnn-nolearn repository with modifications for multiclass segmentation (both iSeg and BraTS).
At some point I might merge the changes. ~However, the final approach (and most likely weights) for the challenge will be public here.~ The approaches contained here are just experimental and not final. They do not represent the approaches we presented to any MICCAI's challenge.