ortho_diff: Plot differences for Prediction and Gold Standard

In neurobase: 'Neuroconductor' Base Package with Helper Functions for 'nifti' Objects

Plot differences for Prediction and Gold Standard

Description

Uses ortho2 to plot differences between a predicted binary image and the assumed ground truth (roi).

Usage

ortho_diff(
  img,
  pred,
  roi,
  xyz = NULL,
  cols = c("#56B4E9", "#D55E00", "#009E73"),
  levels = c("False Negative", "False Positive", "True Positive"),
  addlegend = TRUE,
  center = TRUE,
  leg.cex = 1.5,
  ...
)

multi_overlay_diff(
  x,
  pred,
  roi,
  z = NULL,
  cols = c("#56B4E9", "#D55E00", "#009E73"),
  ...
)

Arguments

img	image to be underlaid
pred	binary segmentation (prediction)
roi	binary manual segmentation (ground truth)
xyz	coordinate for the center of the crosshairs.
cols	colors for false negatives/positives
levels	labels for false negatives/positives
addlegend	add legend, passed to ortho2
center	run xyz on roi. Disregarded if xyz is not NULL
leg.cex	multiplier for legend size
...	arguments to be passed to ortho2 or multi_overlay
x	List of images of class nifti or character vector of filenames
z	slice to display

See Also

ortho2

Examples

set.seed(5)
dims = rep(10, 3)
arr = array(rpois(prod(dims), lambda = 2), dim = dims)
nim = oro.nifti::nifti(arr)
roi = nim > 2
pred = nim > 1.5
ortho_diff(nim, pred, roi)
set.seed(5)
dims = rep(10, 3)
arr = array(rnorm(prod(dims)), dim = dims)
nim = oro.nifti::nifti(arr)
mask = nim > 2
pred = nim > 1.5
multi_overlay_diff(nim, roi = mask, pred = pred)



 if (requireNamespace("brainR", quietly = TRUE)) {
   visits = 1:3
   y = paste0("Visit_", visits, ".nii.gz")
   y = system.file(y, package = "brainR")
   y = lapply(y, readnii)

   y = lapply(y, function(r){
     pixdim(r) = c(0, rep(1, 3), rep(0, 4))
     dropImageDimension(r)
   })

   x = system.file("MNI152_T1_1mm_brain.nii.gz", 
                 package = "brainR")
   x = readnii(x)
   mask = x > 0
   alpha = function(col, alpha = 1) {
       cols = t(col2rgb(col, alpha = FALSE)/255)
       rgb(cols, alpha = alpha)
   }
   roi = y[[2]]
   pred = y
   multi_overlay_diff(x, roi = roi, pred = pred)
   multi_overlay_diff(x, roi = roi, pred = pred, 
         mask = mask, 
         main = paste0("\n", "Visit ", visits),
         text = LETTERS[visits],
         text.x = 0.9,
         text.y = 0.1,
         text.cex = 3)
 }

neurobase documentation built on May 29, 2024, 2:17 a.m.