R/skeletonize.R

In stnava/ANTsR: ANTs in R: Quantification Tools for Biomedical Images

#' Skeletonize a binary segmentation.
#'
#' Employ a topologically constrained label propagation tool in order to thin
#' an existing segmentation into a skeleton.  The function will perform better
#' if the initial segmentation is well-composed.
#'
#' @param x input binary image
#' @param laplacianThreshold between zero and one, controls the speed image
#' @param propagationOption either 0 (none), 1 (well-composed) or 2 (topological) constraint
#' @return output binary image
#' @author Avants BB
#' @examples
#'
#' img <- antsImageRead(getANTsRData("r16")) %>% thresholdImage("Otsu", 3)
#' skel <- skeletonize(thresholdImage(img, 3, 3))
#'
#' @export skeletonize
#'
skeletonize <- function(
    x,
    laplacianThreshold = 0.25,
    propagationOption = 1) {
  # ImageMath 3 wm.nii.gz MD wm.nii.gz 1  ImageMath 3 wm.nii.gz ME wm.nii.gz 1
  #  wm = iMath( x, "MD", 1 ) %>% iMath( "ME", 1)
  wm <- thresholdImage(x, 0, 0) #  ImageMath 3 wm.nii.gz Neg wm.nii.gz
  wmd <- iMath(wm, "D") # ImageMath 3 wmd.nii.gz D wm.nii.gz
  wmdl <- iMath(wmd, "Laplacian", 1.0, 1) # ImageMath 3 wmdl.nii.gz Laplacian wmd.nii.gz 1.0 1
  #  return( wmdl )
  speed <- thresholdImage(wmdl, 0.0, laplacianThreshold) # ThresholdImage 3 wmdl.nii.gz speed.nii.gz 0.0 0.25
  speed <- thresholdImage(speed, 0, 0) # ImageMath 3 speed.nii.gz Neg speed.nii.gz
  wm <- iMath(wm, "GetLargestComponent") # ImageMath 3 wm.nii.gz GetLargestComponent wm.nii.gz
  # ImageMath 3 wm_skel.nii.gz PropagateLabelsThroughMask speed.nii.gz wm.nii.gz 20000 1
  wmskel <- iMath(speed, "PropagateLabelsThroughMask", wm, 200000, propagationOption)
  wmskel <- thresholdImage(wmskel, 0, 0) # ImageMath 3 wm_skel.nii.gz Neg wm_skel.nii.gz
  return(wmskel)
}