3DLAND/Phase I/monai_wholeBody_ct_segmentation/README.md

monai_wholeBody_ct_segmentation

You can run code with your own PC/Notebook/… or on Google Colab.

Create a virtual environment

conda create -n [NAME] python==3.9

Start the environment

conda activate [NAME]

Clone Repository

git clone https://github.com/ytl0623/monai_wholeBody_ct_segmentation.git

Go to the cloned folder

cd monai_wholeBody_ct_segmentation

Install the dependencies

pip install -r requirements.txt

Execute inference

It will cost about three minutes. Check NIFTI directory after run done.

python -m monai.bundle run --config_file configs/inference.json

Unzip inference file

gzip -d NIFTI/DLCSI033/DLCSI033_trans.nii.gz 

Convert NIFTI file to mask file

It will cost about three minutes. Check MONAI directory after run done.

python nii2png.py

Generate DICOM-RT file

It will cost about two minutes. Check DICOM directory after run done.

python main.py

Download DICOM directory

There are two DICOM-RT files. (Original and MONAI)

Download Dicompyler

Show results with Dicompyler

Pay attention to the Chinese path.

Reference

• https://github.com/Kiragroh/Kira_DICOM-RT-Anonymizer-MG
• https://github.com/rordenlab/dcm2niix
• https://github.com/Project-MONAI/model-zoo
• https://monai.io/model-zoo.html