[input volume(s)]

input volume(s)

[xform]

?

[gca file]

?

[output volume]

output volume

-cross-sequence

label a volume acquired with sequence different than atlas

-nogibbs

disable gibbs priors

-wm [path]

use wm segmentation

-conform

interpolate volume to be isotropic 1mm^3

-t1 [gca_t1]

use file to label thin temporal lobe

-debug_voxel [x] [y] [z]

debug voxel

-debug_node [x] [y] [z]

debug node

-debug_label [int n]

debug label

-tr [float n]

set TR in msec

-te [float n]

set TE in msec

-alpha [float n]

set alpha in radians

-example [mri_vol] [segmentation]

use T1 (mri_vol) and segmentation as example

-pthresh [float n]

use p threshold n for adaptive renormalization (default=.7)

-niter [int n]

apply max likelihood for n iterations (default=2)

-novar

do not use variance in classification

-regularize [float n]

regularize variance to be sigma+nC(noise)

-nohippo

do not auto-edit hippocampus

-fwm [mri_vol]

use fixed white matter segmentation from wm

-mri [mri_vol]

write most likely MR volume to mri_vol

-heq [mri_vol]

use histogram equalization from mri_vol

-renorm [mri_vol]

renormalize using predicted intensity values in mri_vol

-flash

use FLASH forward model to predict intensity values

-flash_params [filename]

use FLASH forward model and tissue params in filename to predict

-renormalize [wsize] [iter]

reenorm class means [iter] times after initial label with window of [wsize]

-r [mri_vol]

set input volume

-h

use GCA to histogram normalize input image

-a [int n]

mean filter n time to conditional densities

-w [int n] [filename]

write snapshots of gibbs process every n times to filename

-m [mri_vol]

use mri_vol to mask final labeling

-e [int n]

expand

-n [int n]

set max iterations to n (default=200)

-f [int f] [float t]

filter labeled volume with threshold t (default=.5) mode filter f (default=0)times

output volume

?