This project focuses on segmenting the liver from medical images using a 3D UNet implemented with PyTorch and MONAI. The model is trained and evaluated on the dataset from the Meddical Segmentation Decathlon.

• Introduction
• Installation
• Training
• Demo
• License

Medical image segmentation is a critical task in healthcare, aiding in diagnostics and treatment planning. This project leverages the power of deep learning to accurately segment the liver from 3D medical images.

• Python 3.8+
• CUDA-compatible GPU for acceleration (optional but recommended)

pip install -r requirements.txt
pip install monai

Requirements File

matplotlib
numpy
tqdm
glob2
dicom2nifti
pytest-shutil
nibabel

model = UNet(
    dimensions=3,
    in_channels=1,
    out_channels=2,
    channels=(16, 32, 64, 128, 256), 
    strides=(2, 2, 2, 2),
    num_res_units=2,
    norm=Norm.BATCH,
).to(device)