MRIaggr (version 1.1.3)

calcTissueType: Probabilistic tissue type segmentation

Description

Perform a probabilistic segmentation of the voxel in Cerebro Spinal Fluid, White Matter and Grey Matter classes.

Usage

## S3 method for class 'MRIaggr':
calcTissueType(object, param, niter = 100, nnei = 6,
         beta = if(sub == TRUE){0.3}else{0.7}, sub = TRUE, digit = 0, verbose = TRUE,
         name_newparam = c("CSF", "GM", "WM"), update.object = FALSE, overwrite = FALSE)

Arguments

object

an object of class MRIaggr. REQUIRED.

param

the contrast parameter that should be used to distinguish the WM, the GM and the CSF. character. REQUIRED.

niter

the number of iterations used by mritc.bayes. positive integer.

nnei

the number of neighbors. positive integer.

beta

the parameter 'inverse temperature' of the Potts model. numeric.

sub

if TRUE, use the higher resolution model; otherwise, use the whole voxel method. logical.

digit

the number of decimal places to use for the initialization. positive integer.

verbose

indicate the level of output as the algorithm runs. logical.

name_newparam

the name of the new paramaters containing the probabilistic segmentation. character vector of size 3.

update.object

should the resulting tissue types be stored in object ? logical.

overwrite

if tissue types are already stored in object@data, can they be overwritten ? logical.

Value

• An list containing :
  • [[prob]]: the group membership of each voxel for each class.matrix.
  • [[mu]]: the mean value of each class.numeric vector.
  • [[sigma]]: the standard deviation of each class.numeric vector.

concept

calc.

Details

This function requires to have installed the mritc package to work. ARGUMENTS: Information about the nnei and sub arguments can be found in makeMRIspatial. Information about the niter and beta arguments can be found in mritc. FUNCTION: This function uses the mritc.bayes function of the mritc package to compute the probabilistic segmentation. T1 sequence is the recommanded sequence to identify the various tissue types but T2 gradient echo may also be used. The initialization function initOtsu was found much more slower when the contrast parameter values have a large number of digit. The digit argument enable to round the contrast values ONLY for the computation of the initialization values.

Examples

## load a MRIaggr object
data("MRIaggr.Pat1_red", package = "MRIaggr")

## perform segmentation  (call mritc)
calcTissueType(MRIaggr.Pat1_red, param = "T1_t0", update.object = TRUE, overwrite = TRUE)

## display
multiplot(MRIaggr.Pat1_red, num = 1,
             param = c("CSF","WM","GM"), legend = FALSE,
             palette = "rgb")