CV805 Final Project

Creating the environment

conda env create -f segmentation/environment_droplet.yml

Running segmentation experiments

Determine the model, training strategy and other configurations in segmentation/configs.yaml. To train the segmentation model:

python segmentation/segmentation.py -p segmentation/configs.yaml

Three models are available for training:

  • UNet
  • UNETR
  • SwinUNETR

Results will be saved in the segmentation/results files.