Data Consistency for Magnetic Resonance Imaging
Data Consistency (DC) is crucial for generalization in multi-modal MRI data and robustness in detecting pathology.
This repo implements the following reconstruction methods:
- Cascades of Independently Recurrent Inference Machines (CIRIM) [1],
- Independently Recurrent Inference Machines (IRIM) [2, 3],
- End-to-End Variational Network (E2EVN), [4, 5]
- the UNet [5, 6],
- Compressed Sensing (CS) [7], and
- zero-filled reconstruction (ZF).
The CIRIM, the RIM, and the E2EVN target unrolled optimization by gradient descent. Thus, DC is implicitly enforced. Through cascades DC can be explicitly enforced by a designed term [1, 4].
Installation
You can install mridc with pip:
pip install mridc
Usage
Check on scripts how to train models and run a method for reconstruction.
Check on tools for preprocessing and evaluation tools.
Recommended public datasets to use with this repo:
- fastMRI [5].
Documentation
Read the docs here
License
Citation
Check CITATION.cff file or cite using the widget. Alternatively cite as
@misc{mridc,
author={Karkalousos, Dimitrios and Caan, Matthan},
title={MRIDC: Data Consistency for Magnetic Resonance Imaging},
year={2021},
url = {https://github.com/wdika/mridc},
}
Bibliography
[1] Karkalousos, D. et al. (2021) ‘Assessment of Data Consistency through Cascades of Independently Recurrent Inference Machines for fast and robust accelerated MRI reconstruction’. Available at: https://arxiv.org/abs/2111.15498v1 (Accessed: 1 December 2021).
[2] Lønning, K. et al. (2019) ‘Recurrent inference machines for reconstructing heterogeneous MRI data’, Medical Image Analysis, 53, pp. 64–78. doi: 10.1016/j.media.2019.01.005.
[3] Karkalousos, D. et al. (2020) ‘Reconstructing unseen modalities and pathology with an efficient Recurrent Inference Machine’, pp. 1–31. Available at: http://arxiv.org/abs/2012.07819.
[4] Sriram, A. et al. (2020) ‘End-to-End Variational Networks for Accelerated MRI Reconstruction’, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 12262 LNCS, pp. 64–73. doi: 10.1007/978-3-030-59713-9_7.
[5] Zbontar, J. et al. (2018) ‘fastMRI: An Open Dataset and Benchmarks for Accelerated MRI’, arXiv, pp. 1–35. Available at: http://arxiv.org/abs/1811.08839.
[6] Ronneberger, O., Fischer, P. and Brox, T. (2015) ‘U-Net: Convolutional Networks for Biomedical Image Segmentation’, in Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention, pp. 234–241. doi: 10.1007/978-3-319-24574-4_28.
[7] Lustig, M. et al. (2008) ‘Compressed Sensing MRI’, IEEE Signal Processing Magazine, 25(2), pp. 72–82. doi: 10.1109/MSP.2007.914728.
![[BUG] Check if half_scan_percentage alters the chosen acceleration factors / Add option for 1D masks](https://avatars.githubusercontent.com/u/22858443?v=4)
![[pre-commit.ci] pre-commit autoupdate](https://avatars.githubusercontent.com/in/68672?v=4)
89 Dec 09, 2022
1.4k Jan 08, 2023
167 Jan 02, 2023
336 Nov 25, 2022
5 Dec 02, 2022
127 Dec 16, 2022
828 Jan 04, 2023
32 Dec 21, 2022
3 Aug 03, 2022
137 Sep 10, 2022
1.3k Jan 02, 2023
55 Dec 12, 2022
65 Dec 14, 2022
470 Dec 30, 2022
28 Dec 30, 2022
6 Nov 09, 2021
10 Sep 20, 2022
214 Dec 29, 2022
368 Dec 17, 2022
155 Dec 26, 2022