This code solves the optimization problem of water/fat separation in MRI by reconstructing water, fat, field, and R2* from input complex GRE data with 6 echoes using No-Training Deep Neural Network (NTD) method. For details see: http://dx.doi.org/10.1002/mrm.28546
Download input_data.mat: https://wcm.box.com/s/b1fiikn3shqy44v3eewfia8unit9ybse
To run the code, download all the files into a folder, make sure all the libraries listed in main.py exist and run:
CUDA_VISIBLE_DEVICES=0 python main.py input_data.mat output_data.mat
To plot results in MATLAB run run_results.m