This repository provides a pytorch implementation of xSDNet, as presented in our paper "Joint image reconstruction and segmentation of real-time cardiac MRI in free-breathing using a model based on disentangled representation learning, JCMR, 2025".
The code is based on and derived from Spyridon Thermos' implementation of the SDNet model as used in this study. The Spatial Decomposition Network (SDNet) itself is introduced here. To use the model, you need to copy the layers folder from Spyridon's repo, as those implementations are imported in xSDNet.py.