Lungs-Tumour-segmentation

This project aims to segment lungs tumor from CT scans of Decathlon Lungs dataset using Pytorch and MONAI .

This project is still under development.

Directory structure

├── LICENSE                    # License file
├── README.md                  # Readme file
├── app
│   ├── gunicorn.py            # Configuration for Gunicorn server
│   ├── main.py                # Main application file
│   └── schemas.py             # Schema definitions
├── artifacts
│   └── checkpoints            # Directory for storing model checkpoints
├── conf
│   └── config.yaml            # Configuration file in YAML format
├── data/                      # data dir
├── logs/                      # Log dir  
├── notebook.ipynb             # Jupyter notebook file
├── requirements.txt           # File listing required Python packages
├── src/
│   ├── data_module.py         # Module for data processing
│   ├── model.py               # Module containing the model definition
│   └── utils.py               # Utility functions module
├── download_data.sh           # Shell script for downloading data
└── train.py                   # Script for training the model

Workflows

1. Set up environment.

python -m venv venv         # create environment
source venv/bin/activate    # activate environment

2. Train model

python train.py

References

• PyTorch
• MONAI
• LightningAI
• Brain Tumor Segmentation using MONAI and WandB
• https://github.com/Project-MONAI/tutorials/tree/main/3d_segmentation

Authors

• @bhimrazy

License

MIT