Download dataset

To download full 40GB dataset follow this:

1. Download 'archive.zip' from https://www.kaggle.com/datasets/washingtongold/lidcidri30
2. Unzip it into lidc directory (this could take a while)

unzip archive.zip -d lidc

Run prepare data script to convert the dataset into a format that can be used by the model.

Our prepare-data.py script uses pylidc which uses some old dependencies. With python 3.9 everything works fine. But for model training/inference you should use latest python version.

python3.9 -m venv venv
source venv/bin/activate
python3.9 -m pip install -r prepare-data-requirements.txt
python3.9 prepare-data.py

Or you can use already prepared zip of 3000 images to save time data.zip.

References

• Road Extraction by Deep Residual U-Net
• ResUNet-a: a deep learning framework for semantic segmentation of remotely sensed data
• ResUNet++: An Advanced Architecture for Medical Image Segmentation