DeepSTI: Towards Tensor Reconstruction using Fewer Orientations in Susceptibility Tensor Imaging

This is the official implementation of the paper

DeepSTI: Towards Tensor Reconstruction using Fewer Orientations in Susceptibility Tensor Imaging. Medical Image Analysis 2023.

by Zhenghan Fang, Kuo-Wei Lai, Peter van Zijl, Xu Li, and Jeremias Sulam.

Requirements

• Python 3.9
• PyTorch 1.12.0

Environment Settings

Use the command below to install all required libraries.

conda env create --name [MY_ENV] -f environment.yml

Usage

Activate conda environment first

conda activate [MY_ENV]

Train

python deepsti/main.py 

arguments:
--mode                        train (train or predict)
--name                        name of your experiment
--data_dir                    path to dataset directory
--train_list                  list of training data
--validate_list               list of validation data
--test_list                   list of testing data
--tesla                       field strength in training data [default: 3]
--batch_size                  batch size [default is 2]
--gpu                         GPU ID's, e.g. "0" or "0,1"

Example:

python deepsti/main.py --mode train --name myexp --data_dir data/ --train_list train.txt --validate_list validate.txt --test_list test.txt --gpu 0,1

Tensorboard Visualization

tensorboard --logdir experiment/tb_log/deepsti_resunet_myexp

Test on External Data

python deepsti/main.py

arguments:
--mode                        predict (train or predict)
--resume_file                 saved model parameters
--ext_data                    yml file of external data information
--gpu                         GPU ID's, e.g. "0" or "0,1"

Example:

python deepsti/main.py --mode predict --resume_file experiment/checkpoint/deepsti_resunet_Vmodel.pkl --gpu 1 --ext_data data/yml/example.yml

Predictions will be saved in experiment/results.

Dataset

Demo data will be provided shortly.

References

If you find the code useful for your research, please consider citing

@article{fang2023deepsti,
  title={Deepsti: towards tensor reconstruction using fewer orientations in susceptibility tensor imaging},
  author={Fang, Zhenghan and Lai, Kuo-Wei and van Zijl, Peter and Li, Xu and Sulam, Jeremias},
  journal={Medical image analysis},
  volume={87},
  pages={102829},
  year={2023},
  publisher={Elsevier},
  doi={https://doi.org/10.1016/j.media.2023.102829}
}