Skip to content
/ LOUPE Public
forked from cagladbahadir/LOUPE

Learning-based Optimization of the Under-sampling Pattern in MRI

Notifications You must be signed in to change notification settings

VC86/LOUPE

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

50 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

LOUPE

Learning-based Optimization of the Under-sampling PattErn

Python implementation in keras/tensorflow of LOUPE, which simultanously optimizes the under-sampling pattern and reconstruction model for MRI. See abstract and paper for more details.

Training

Please see train_ipmi.py for an example of how to train LOUPE (this uses FASHION_MNIST as an example dataset).

Citation

If you use the open source code, please cite:

  • Deep-learning-based Optimization of the Under-sampling Pattern in MRI
    C. Bahadir‡, A.Q. Wang‡, A.V. Dalca, M.R. Sabuncu.
    IEEE TCP: Transactions on Computational Imaging. 6. pp. 1139-1152. 2020. arXiv:1907.11374.

  • Learning-based Optimization of the Under-sampling Pattern in MRI.
    Cagla D. Bahadir, Adrian V. Dalca, and Mert R. Sabuncu.
    IPMI: Information Processing in Medical Imaging. 2019. arXiv:1901.01960.

Legacy Code (v1.0)

Code for the original LOUPE code was moved to the legacy folder.

Abstract

Learning-based Optimization of the Under-sampling Pattern in MRI Acquisition of Magnetic Resonance Imaging (MRI) scans can be accelerated by under-sampling in k-space (i.e., the Fourier domain). In this paper, we consider the problem of optimizing the sub-sampling pattern in a data-driven fashion. Since the reconstruction model's performance depends on the sub-sampling pattern, we combine the two problems. For a given sparsity constraint, our method optimizes the sub-sampling pattern and reconstruction model, using an end-to-end learning strategy. Our algorithm learns from full-resolution data that are under-sampled retrospectively, yielding a sub-sampling pattern and reconstruction model that are customized to the type of images represented in the training data. The proposed method, which we call LOUPE (Learning-based Optimization of the Under-sampling PattErn), was implemented by modifying a U-Net, a widely-used convolutional neural network architecture, that we append with the forward model that encodes the under-sampling process. Our experiments with T1-weighted structural brain MRI scans show that the optimized sub-sampling pattern can yield significantly more accurate reconstructions compared to standard random uniform, variable density or equispaced under-sampling schemes.

Trained Model Weights

Trained model weights for LOUPE and individual U-Nets for different mask configurations are available upon request due to large file sizes. Please contact Cagla Deniz Bahadir ([email protected]) for the weight files.

About

Learning-based Optimization of the Under-sampling Pattern in MRI

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%