Many thanks to the Sunnybrook Health Sciences Centre for providing a set of CMR data with associated contours. Unfortunately, in the latest release the filenames have become a little mangled, and don't match up with the contours. I have gone through the files and matched them up; exported the DICOMS as PNGs and converted the list of coordinates of the contours to PNGs as well.
The first two sets of CMRs are included as training data, the last set as test data.
The original SegNet uses max_pool_with_argmax, and requires an unpool_with_argmax. Unfortunately, Tensorflow does not provide an unpool_with_argmax. Fortunately there is code in the github thread above to make your own.
This version of unpool_with_argmax runs on the CPU not GPU so is a little slower.
Tensorflow does not provide a CPU version of max_pool_with_argmax, so if you don't have a GPU you can't run this.
Tensorflow forgot to include a function for gradients for maxpoolwithargmax, so it is included at the bottom of train.py
The name mangling of the Sunnybrook CMR data - I have fixed this and the data is included in the download.
SegNet works better with a version of softmax that is inversely weighted by class frequency.
Now using SELU as the activation funciton - this allows us to get rid of the Batch Norm (and the associated is_training hassle).