A Pytorch-Version for DREAM-Net. Some details are different from the original paper.
- Install torch-randon (https://github.com/matteo-ronchetti/torch-radon)
- Generate simulated sparse-view CT data: gen_ld_data.py
- Generate training data: make_proj_img_list.py
- Train and Validation: train_dreamnet.py
- Test: test_dreamnet.py
Please cite the following references:
- TorchRadon: Fast Differentiable Routines for Computed Tomography
- DREAM-Net: Deep Residual Error iterAtive Minimization Network for Sparse-View CT Reconstruction