CNN Segmentation Package

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.

Submodules:

Segmentation

• segmentation.callbacks: callbacks used during the training of the network
• segmentation.cnn: this module allows to define CNN-based models in a dynamic way
• segmentation.data: this module implements functions to handle data in various way
• segmentation.losses: this module allows to define loss functions to be used for segmentation tasks
• segmentation.metrics: in this module, metrics to be used during the training of the network are implemented
• segmentation.utils: utils functions used for various purposes

Segmentation.preprocess

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 functions
• segmentation.preprocess.Cycles: this module contains example functions to loop over files and call the main functions
• segmentation.preprocess.utils: utils functions used for various purposes

Current lab students working on the project: - Alberto Faglia: AlbiFag

Contribution

When contributing to the code and before submitting a pull request, please make sure to run make -B to update the documentation of the project.

Documentation

Documentation of the package can be found at the following link: https://ltebs-polimi.github.io/cnn-segmentation/