The algorithms presented here are described in: "Presentation of microstructural diffusion components by color schemes in abdominal organs" in Magnetic Resonance in Medicine (Article DOI: 10.1002/mrm.30183). When using RGBtri, please cite the paper.
RGBtri is a tool for the creation of color-coded maps of the diffusion components of triexponential diffusion-weighted imaging (DWI). The following tri-exponential model is assumed.
S(b)/S0 = Fs e-bDs + Fi e-bDi + Ff e-Df.
RGBtri merges the diffusion fractions of slow, intermediate, and fast components int one rgb image.
The inputs to RGBtri are maps of the fast (Ff) and intermediate (Fi) diffusion components. (Ff and Fi should be scaled as a percentage so that their values range from 0 to 100. If Ff and Fi are between 0 and 1, use the "-f" option when running RGBtri.) The slow component (Fs) is calculated internally by the equation Fs = 1 - Fi - Ff. The output of RGBtri is a color-coded parametric image, where Ff uses the red, Fi the green, and Fs the blue color channel.
The syntax of RGBtri is quite simple: "RGBtri -Ff file1.nii -Fi file2.nii -o output.nii". Type "RGBtri -h" for more details. Some test data is available in the "test-data" directory. A Windows executable is provided in the "win-binary" directory. A Linux executable (Ubuntu 22.04.4 LTS) can be found in the "ubuntu-binary" directory.
RGBtri is written in C++. A cmake file has been added for easy compilation on different operating systems. Since RGBtri software works with Nifti images, the only dependency is the Nifti package: https://github.com/NIFTI-Imaging/nifti_clib