Recurrent Mask Refinement for Few-Shot Medical Image Segmentation

Steps

1. Install any missing packages using pip or conda

2. Preprocess each dataset using utils/preprocess

3. Run notebooks/prepare_data_for_few_shot_learning.ipynb to generate files containing the range of each organ

Train

Test

Configure the yaml. An example is at yamls/example.yml. Change ckpt to trained model checkpoint python test.py test --yaml $PATH_TO_YAML