Skip to main content
Ctrl+K
bartorch bartorch
  • Homepage
  • User guide
  • Developer guide
    • Development installation
    • Development workflow
    • Coding style
    • Documentation
    • Terminology and conventions
    • Internals
      • Array layouts and file conventions
      • Encoding form and lowering
      • NUFFT substitution
      • Regularized least squares
      • Model-based command routing
      • Backends and timings
    • Repository layout
    • BART fork
  • Explanations
    • Execution model
    • Data layout
    • Encoding operators
    • Non-Cartesian transform
    • Regularizers
    • Solvers
    • Model-based reconstruction
    • Learned reconstruction
    • Performance
  • Examples
    • Course
      • 1. Your data as tensors
      • 2. From k-space to an image
      • 3. Coil sensitivities
      • 4. Joint reconstruction of image and sensitivity maps
      • 5. Choosing a regularizer
      • 6. Custom encoding and solver
      • 7. Basics of non-Cartesian imaging
      • 8. Radial SENSE
      • 9. Nonlinear inversion for T2 mapping
      • 10. Plug-and-play reconstruction
      • 11. Training an unrolled network
    • Tours
      • Readout oversampling and apodization
      • EPI Nyquist ghost and ramp sampling
      • Receive bias field
      • Gradient nonlinearity
      • Off-resonance correction of spiral imaging
      • Rigid head motion from navigators
      • Susceptibility distortion in EPI
      • Noise prewhitening
      • Dynamic golden-angle radial MRI
      • Subspace-constrained T1 mapping
      • Parameter maps from scanner images
      • Networks for complex volumes
      • Staged training of an unrolled network
      • Training without a reference
      • Annealed plug-and-play
      • Uncertainty estimation
  • API reference
    • Array functions and settings
    • BART commands as functions
    • Linear operators
    • Nonlinear operators
    • Optimization
    • Regularization and denoising
    • Reconstruction pipelines
    • Learning
    • Interoperability
    • File I/O
    • Command line
  • Miscellaneous
    • License and third-party notices
    • Related projects
    • Contributors and citation
  • Repository
  • Suggest edit
  • Open issue
  • .md

Related projects

Contents

  • Components of bartorch
  • Other reconstruction frameworks

Related projects#

Components of bartorch#

Project

Relationship

BART

The reconstruction toolbox bartorch embeds, developed at codeberg.org/mrirecon/bart and pinned as the external/bart submodule from the downstream fork pulserver/bart

PyTorch

Tensors, devices and automatic differentiation of every interface

FINUFFT and cuFINUFFT

Non-uniform fast Fourier transforms substituted for BART’s gridding on the host and on CUDA devices

MRI-NUFFT

Fits the time-segmentation coefficients of bartorch.linop.FieldCorrected()

BlochSim

Signal simulation for the signal models of bartorch.nlop

DeepInverse

Target of bartorch.interop.to_deepinv(), and a source of denoisers for ImplicitPrior

SimpleITK

Resampling for bias field and gradient nonlinearity correction, and registration of navigator planes, in bartorch.tools

PyHySCO

Susceptibility distortion correction behind bartorch.tools.correct_susceptibility()

Other reconstruction frameworks#

Project

Relationship

SigPy

Python operators, proximal operators and MRI reconstruction apps, implemented in NumPy and CuPy; an independent implementation for checking conventions

MRpro

MRI reconstruction implemented in PyTorch itself, with its own operators and data handling

MRIReco.jl

MRI reconstruction in Julia

Pyxu

General computational-imaging framework; its treatment of forward operators, functionals and proximal algorithms is the conceptual reference for Regularizers and Solvers

previous

License and third-party notices

next

Contributors and citation

Contents
  • Components of bartorch
  • Other reconstruction frameworks

By bartorch contributors

© Copyright 2024–2026, bartorch contributors.