/JMONSEL-inverse

Primary LanguagePython

JMONSEL-inverse

  • Please fork the github link : https://github.com/saadullah01/JMONSEL-inverse/
  • AUTOMAP: the updated code of AUTOMAP for our problem.
  • data: contains output of JMONSEL (train_input, test_input) and ground truth (train_x_real, test_x_real).
  • fine_tuned_model/checkpoint: weights for fine-tuned AUTOMAP model with training part of JMONSEL output and ground truth images provided in data folder.
  • pre_trained_model/checkpoint: weights for pre-trained AUTOMAP model with MRI Training and Testing Data.

To Run

First you need to change configuration in AUTOMAP/configs/inference_64x64_ex.json. Give loadmodel_dir the path to the fine_tuned_model/checkpoint and add path for the data directory to data_dir.

Run the following command to create inferrence of the AUTOMAP output

python AUTOMAP/automap_main_inference.py -c AUTOMAP/configs/inference_64x64_ex.json

JMONSEL-CODE

  • More information about JMONSEL along with permission for it's code is given JMONSEL-NIST
  • The folder jmonsel-code contains the output files of JMONSEL along with the code used to define the materials Silicon and Glassy Carbon. The code-file image-constructor.py was used to extract the information out of the textfiles generated as output of JMONSEL and converted it to a numpy matrix used for further processing.