LAUnet

Deep learning program for left atrial and scar segmentation in GE-MRI images.

File and directory structure

• artifical_data/: Generate artificial scar in images
• core/: Main scripts of the repository
  • architectures/: different network architectures
  • augmentations/: scripts for offline and online augmentation
  • helper_functions.py: helper functions with general functions that are useful through the repository
  • imshow_3D.py: function to show 3D images
  • predict.py: predict a single image using a trained network
  • settings.py: file with main settings object
  • test.py: test the trained network
  • train.py: train a network
• data_exploration/: files for exploration of the data
• procedures/: scripts that execute training and or testing procedures using files in the core directory
• visualisations/: script for visualising results