PyTorch implementation for our paper (submitted):
"Deep matched filtering for retinal vessel segmentation"
The full material will be made available once the manuscript is accepted!!!
Project for WS-DMF
├── data
├── eyeset.py (dataset & dataloader & data pre-processing)
└── ...
├── nets
├── modules/activation.py (activation functions)
├── modules/attention.py (attention modules)
├── conv.py (convolution layers)
├── dmfu.py (deep matched filtering)
├── lunet.py (lightweight UNet)
├── rot.py (APC layer related OAL loss function)
└── ...
├── utils
├── loss.py (loss function)
├── optim.py (optimizer for training)
└── ...
├── build.py (implementation for WS-DMF)
├── grad.py (implementation for backgrading)
├── loop.py (implementation for training)
├── main.py (implementation for main function)
└── ...
├── onnx (Pytorch trained weights)
└── infer.py (infer fundus images for segmentation with *.onnx weights)
└── ...
├── results (segmentation results)
└── ... (segmentation results for popular datasets)
└── ... (segmentation results for cross-dataset-validation)
For any questions, please contact me. And my e-mails are