ROI selection

ROI types

There are four types of ROI that can be used for tracking:

seed
the tracking algorithm will initiate tracks from random location within these regions.
include
target ROIs: tracks generated by the algorithm will need to pass through these ROIs; tracks that do not enter all of these ROIs will be discarded.
exclude
tracks that enter any of these ROIs will immediately be discarded.
mask
tracks must be contained within these ROIs; track will be terminated (but not discarded) as soon as they leave the ROI.

top

Defining ROIs

ROIs can be specified as spherical regions, or as a mask image.

Spherical ROIs

These are best defined using MRView. For example: 

> mrview anat.mif

Set MRView to display cross-hairs on the current focus, using the 'View->Show Focus' menu:

Place the cross-hairs in the desired seed location, and read off the real-world position from the statusbar, as shown below:

The ROI can then be specified using this position, along with its desired radius. For convenience, the position can be dropped directly onto the command-line using drag & drop from the statusbar.

Mask ROIs

Regions of interest can also be specified in the form of mask images. In this case, the set of voxels with intensities greater than 0.5 constitutes the ROI. The 'ROI analysis' of MRView can be used to define such ROIs (see ROI analysis for details).
top

Donald Tournier
MRtrix version 0.2.12
Last updated 2012-01-06