This repository contains the source code of the segmentation package. This package provides utility functions to be used for CNN-based segmentation of medical images.
segmentation.callbacks
: callbacks used during the training of the networksegmentation.cnn
: this module allows to define CNN-based models in a dynamic waysegmentation.losses
: this module allows to define loss functions to be used for segmentation taskssegmentation.metrics
: in this module, metrics to be used during the training of the network are implementedsegmentation.utils
: utils functions used for various purposes
Submodule to preprocess dicom or Nifti files to adapt data for training and optimize memory
segmentation.preprocess.main_preprocess
: this module contains the main preprocessing functionssegmentation.preprocess.Cycles
: this module contains example functions to loop over files and call the main functionssegmentation.preprocess.utils
: utils functions used for various purposes
Current lab students working on the project: - Alberto Faglia: AlbiFag
You can install the package directly from GitHub:
pip install +git https://github.com/ltebs-polimi/cnn-segmentation
You can generate your local documentation by running: make docs
to update the documentation of the project. Updated documentation can be found in the docs
folder.
Documentation of the package can be found at the following link: https://ltebs-polimi.github.io/cnn-segmentation/