Gaussian smooth image using FSL

Source: R/fsl_smooth.R, R/fslhd.R

This function calls fslmaths -s to smooth an image and either saves the image or returns an object of class nifti

fsl_smooth(..., outfile = tempfile(fileext = ".nii.gz"), retimg = FALSE)

fslsmooth(
  file,
  sigma = 10,
  mask = NULL,
  smooth_mask = TRUE,
  smoothed_mask = NULL,
  outfile = NULL,
  retimg = TRUE,
  reorient = FALSE,
  intern = FALSE,
  verbose = TRUE,
  ...
)

Arguments

...

additional arguments passed to readnii.
outfile

(character) resultant smoothed image name (optional) if not give, will be the stub of the filename then _sigma
retimg

(logical) return image of class nifti
file

(character or nifti) image to be smoothed
sigma

(numeric) sigma (in mm) of Gaussian kernel for smoothing
mask

(character) optional mask given for image
smooth_mask

(logical) Smooth mask? If TRUE, the masked image will be divided by the smoothed mask.
smoothed_mask

(character or nifti) If specified and smooth_mask = TRUE, then will use this as the smoothed mask for division.
reorient

(logical) If retimg, should file be reoriented when read in? Passed to readnii.
intern

(logical) to be passed to system
verbose

(logical) print out command before running

Value

Result from system command, depends if intern is TRUE or FALSE.

Note

Functions with underscores have different defaults and will return an output filename, so to be used for piping

Examples

if (have.fsl()){
system.time({
dims = c(50, 50, 20)
x = array(rnorm(prod(dims)), dim = dims)
img = nifti(x, dim= dims, 
datatype = convert.datatype()$FLOAT32, cal.min = min(x), 
cal.max = max(x), pixdim = rep(1, 4))
s.img = fslsmooth(img, retimg=TRUE)
})
}