BSEResU-Net

The Tensorflow-keras Implementation of the BSEResU-Net: An attention-based before-activation residual U-Net for retinal vessel segmentation

Requirements

• Tensorflow 2.2.0+
• Python 3.6+
• PIL
• scikit-learn
• opencv
• h5py

Usage

Training

• Install this repository and the required packages.

• Prepare dataset

  • Download DRIVE dataset at the the official website.
  • Create HDF5 datasets of the ground truth, masks and images for both training and testing.

    python prepare_datasets_DRIVE.py
    

• specify configuration in the file configuration.txt.

• Train BSEResU-Net

  python run_training.py
  

Testing

python run_testing.py

Results

Citation

@article{li2021bseresu,
  title={BSEResU-Net: An Attention-based Before-activation Residual U-Net for Retinal Vessel Segmentation},
  author={Li, Di and Rahardja, Susanto},
  journal={Computer Methods and Programs in Biomedicine},
  pages={106070},
  year={2021},
  publisher={Elsevier}
}

Acknowledgments

This code is built based on retina-unet and we modified it to be compatible to Tensorflow 2.2+.

License

This project is licensed under the MIT License