This is Keras implementation of a fully convolutional neural network with residual connections for automatic segmentation of prostate structures from MR images.
More info on this competition can be found on Grand Challenges website. Data can be downloaded from https://promise12.grand-challenge.org/download/
The network architecture was inspired by U-Net: Convolutional Networks for Biomedical Image Segmentation and by Keras implementation of the model by Paul-Louis Pröve.
The predictions of this model achieved the score 83.70 and was ranked #8 in the competition. For more details about the model and the implementation see the file project_summary.pdf
This tutorial depends on the following libraries:
- scikit-image, numpy, matplotlib, scipy
- SimpleITK
- OpenCV
- Tensorflow >=1.4
- Keras >= 2.0
This code should also be compatible with Theano backend of Keras, but in my experience Theano is slower than TensorFlow.
-
Run
python train.pyto pre-process the data and train the model. Model weights are save in file../data/weights.h5. -
Run
python test.pyto test the model on the train and validation set and generate some images with some best and worst predictions.