- Ubuntu 20.04 LTS
- RTX 3090 with 24GB GPU memory
- 11.7
- Python 3.10
To install requirements:
git clone https://github.com/luckieucas/FLARE23.git
cd nnUNet
pip install -e .
- MICCAI FLARE 2023 dataset
- cropping
- intensity normalization
- resampling
- flip
- rotation
- scale
Running the data preprocessing code:
nnUNetv2_plan_and_preprocess -d 12 --verify_dataset_integrity
- To train the model(s) in the paper, run this command:
python run_training_Flare.py 12 3d_mylowres 1 -tr nnUNetTrainerFlarePseudoCutUnsupLow -p nnUNetPlans
- To infer the testing cases, run this command:
nnUNetv2_predict -i <path_to_data> -o <path_to_output_data> -d 12 -c 3d_mylowres -f 1 -chk <name_of_trained_model> -tr nnUNetTrainerFlarePseudoCutUnsupLow -step_size 0.6 -npp 3 --disable_tta
To compute the evaluation metrics, run:
nnUNetv2_evaluate_folder <path_to_ground_truth> <path_to_inference_results>
Our method achieves the following performance on MICCAI FLARE23: Fast, Low-resource, and Accurate oRgan and Pan-cancer sEgmentation in Abdomen CT
Model name | Organ DICE(%) | Organ NSD(%) | Tumor DICE(%) | Tumor NSD(%) |
---|---|---|---|---|
Our method | 92.18 | 96.33 | 46.26 | 38.65 |
Pick a license and describe how to contribute to your code repository.
We thank the contributors of public datasets.