Spatial Decomposition Layer Segmentation network.
This repository contains a reference PyTorch implementation of SD-LayerNet, presented in SD-LayerNet: Semi-supervised retinal layer segmentation in OCT using disentangled representation with anatomical priors.
The source-code is currently provided as a reference, and cannot be trained out of the box, however, it contains the full implementation of SD-LayerNet (in the folder architecture), as well es the training code in model.py.
- PyTorch 1.9.1
- Cuda 11
- Python 3.9
- opencv-python
Soon we will add code for training on public datasets.