'Radiomic' Image Processing Toolbox

Functions to extract first and second order statistics from images.

The radiomics package is a set of tools for computing texture matrices and features from images.

The release version of this package (April 2016, v0.1.2) is available from CRAN using:


Or you can install the development version of the package using:


Texture Matrices

In the package are functions for calculating four different types of matrices and associated feature sets used to quantify the texture of an image.

These matrices are the:

  • Grey Level Co-occurrence Matrix
  • Grey Level Run Length Matrix
  • Grey Level Size Zone Matrix
  • Multiple Grey Level Size Zone Matrix

Detailed descriptions and references for each matrix can be found here. Detailed usage directions for calculating features and matrices can be found in the package vignette (use browseVignettes(package = "radiomics"))

Using the Package

Texture matrices can be created from 2D images by using the abbreviated and lowercase matrix name as a function call:

tumor <- radiomics::tumor #2D MRI slice of a brain tumor

A matrix with the class of the texture matrix type is returned, as shown here using glcm(tumor, n_grey=4) #> 1 2 3 4 #> 1 0.1617021277 0.03356974 0.001891253 0.0004728132 #> 2 0.0335697400 0.38345154 0.010638298 0.0014184397 #> 3 0.0018912530 0.01063830 0.301654846 0.0184397163 #> 4 0.0004728132 0.00141844 0.018439716 0.0203309693

class(glcm(tumor, n_grey=4))[1]
#> [1] "glcm"

Each matrix type has an associated image function for visualization of the results:


The image functions make use of the viridis scale, as shown here using image(glcm(tumor, n_grey=64)):

Each matrix type has an associated calc_features function, which returns an object of class data.frame with a single observation for each calculated feature. First order features can also be calculated on 2D matrices.



radiomics 0.1.2

  • Numerous bugfixes/edge cases
  • Better handling of NAs in matrices
  • Minor optimizations

radiomics 0.1.1

  • Added generator functions for classes
  • Added option to glrlm, glszm and mglszm truncate output matrix
  • Changed algorithm used to calculate glcm, resulting in speed increases for larger images

radiomics 0.1.0

  • First release!

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.


0.1.3 by Joel Carlson, a month ago

Browse source code at https://github.com/cran/radiomics

Authors: Joel Carlson

Documentation:   PDF Manual  

MIT + file LICENSE license

Imports spatstat, reshape2, methods, Rcpp

Suggests testthat, knitr, rmarkdown, viridis, devtools, roxygen2

Linking to Rcpp

See at CRAN