/STFR-MWF

Code for reproducing results in "Myelin Water Fraction Estimation Using Small-Tip Fast Recovery MRI".

Primary LanguageJulia

STFR-MWF

This repository contains code for reproducing the results in the paper:

Steven T. Whitaker, Gopal Nataraj, Jon-Fredrik Nielsen, and Jeffrey A. Fessler.
Myelin water fraction estimation using small-tip fast recovery MRI.
Magnetic Resonance in Medicine. 2020;00:1-14.
https://doi.org/10.1002/mrm.28259

Getting Started

The code in this repository is structured as a Julia module. It was coded using Julia 1.1.1, though other versions may work as well.

  1. Download or clone this repository.
  2. Download the BrainWeb dataset here.
    • Choose raw byte (unsigned) for the file format.
    • Choose none for compression.
    • Place the downloaded file in <path_to_this_repo>/STFR-MWF/src/estimation/data/BrainWeb/, where <path_to_this_repo> is where you downloaded this repo on your computer. (You will have to create the data and BrainWeb directories.)
  3. Download Julia 1.1.1 here.
  4. Run Julia.
  5. Change directories to this repo with 
    julia> cd("<path_to_this_repo>")
  6. Install some necessary packages and load the code by running setup.jl via 
    julia> include("setup.jl")
  7. Run any function with 
    julia> STFRMWF.func() # Replace func with actual function name

Steps 1 through 3 only need to be done once, and steps 4 through 6 only need to be done once each time you start Julia. (Note that running setup.jl again will not reinstall packages, so will run faster after the first time.)

Reproducing Results

To reproduce the results in the paper, call the corresponding function. For example, to reproduce Figure 3 run

julia> STFRMWF.figure3()

Some results are used for both a figure and a table. For example, Figure 4 and Table 3 use the same data, so to reproduce both run

julia> STFRMWF.figure4table3()

To get a list of all available functions, run

julia> STFRMWF.reproducible

The first time calling these functions may take a while (especially the biased CRLB and NNLS results), but after running once the results will be saved, so future calls will simply load the results and display them.

Data

The data used in the paper and in this code is available at Deep Blue Data.