/ADPKD

Primary LanguagePython

Download docker image

$ docker pull piotrekwoznicki/adpkd-net:v0.1

More instructions on the inference can be found at Docker Hub.

Run

1. Launch the container:

$ bash launch_container.sh [path_to_dataset - will be mounted at /workspace/data]

2. Run inference:

$ cd /workspace/source
$ bash fit.sh  \
    -i / --InputVol [abs_path_to_nifti_image]
    -o / --OutputDir [where_to_save_segmentation_and_JSON]
    --small [whether to use single model - OPTIONAL (requires shorter calculation time)]

3. Run inference for test:

$ cd /workspace/source
$ bash fit.sh  \
    -i /workspace/test_data/T2_ax.nii.gz
    -o /workspace/data/test_results_ax

$ bash fit.sh  \
    -i /workspace/test_data/T2_cor.nii.gz
    -o /workspace/data/test_results_cor

Before building the docker image, download the folder trained_models from GDrive and place it in the source directory