Skip to content
This repository has been archived by the owner on Apr 25, 2022. It is now read-only.

Latest commit

 

History

History
49 lines (40 loc) · 1.29 KB

README.md

File metadata and controls

49 lines (40 loc) · 1.29 KB

Binder

Online MRI Reconstruction

Installation

Dependencies:

  • PySAP
  • PySAP-mri
  • ModOpt
  • And their dependencies.
  1. Create a virtual environment (Python 3.6+)
  2. Ensure to have OpenCL/ Cuda installed and setup on your machine
sudo apt install nfft # required for pynfft
pip install -r requirements.txt 
pip install --upgrade git+https://github.com/CEA-COSMIC/pysap.git
pip install --upgrade git+https://github.com/paquiteau/pysap-mri.git

TODO:

Parameters and Hyper-parameters

Inputs parameters

  • Mono- vs Multi-coil MRI
  • Size of the reconstruction
  • Size of the kspace acquision
  • Acquisition strategy of the kspace

Reconstruction Operator

  • Fourier Operator: FFT2D, FFT1.5D, NUFFT
  • Regularisation Domain: Wavelet(type, nb_scale)...
  • Regularisation function:
    • Norm l_1 l_2 l_{1,2}, OWL
    • Order of the dimension (per coil, per bands, all coef...)
  • Gradient formulation: Analysis vs Synthesis

Reconstruction Algorithm

Forward-Backward

-> Regularisation at which step

  • POGM
  • FISTA
  • Vanilla Gradient Descent
  • Custom Gradient Descent Strategies

Primal-Dual

-> regularisation mandatory at each step (use of the dual variable)

  • Condat Vu