nnUNet-package 0.2.4

Package to use nnUNet on 3D Slicer

LungSegmentation nnUNetv2 Prediction Script

A Python script to easily run nnUNetv2 lung segmentation predictions on medical images. Automatic model download, file preparation for prediction, and result renaming included.

---

Features

• Automatic download and extraction of models from a URL.
• Preparation of dataset.json for nnUNet prediction.
• Conversion of input images to .nrrd if necessary.
• Prediction execution with detailed logs.
• Automatic cleanup of temporary files.
• Automatic renaming of the final prediction file.

---

Requirements

Before running the script, make sure you have installed and configured the following:

git clone https://github.com/FlorianDAVAUX/nnUNet_package.git
cd nnUNet_package
pip install -e .

---

Usage

---

Option Description Example

---

--mode Prediction mode --mode Invivo (Invivo or Exvivo)

--structure Structure to segment --structure Parenchyma (Parenchyma,
Airways, Vascular,
ParenchymaAirways,
All, Lobes)

--input Path to the input image --input ~/data/scan_patient.nrrd (.nii, .nii.gz, .mha,
.nrrd)

--output Output directory for the --output ~/predictions prediction (default:
prediction)

--models_dir Path to store or search --models_dir ~/models for models

--name Final name of the --name segmentation_parenchyma prediction file (without extension)

---

Full Example

nnunet_predict     --mode Invivo     --structure Parenchyma     --input ~/data/scan_patient.nrrd     --output ~/predictions     --models_dir ~/models     --name segmentation_parenchyma