crispify

create a crisp labelled volume from multiple fuzzy classification volumes

crispify <options> <infile> <class> ...

DESCRIPTION

crispify takes multiple fuzzy (probabilistic) classification volumes as input, each paired with an integer class label, and produces a single crisp (discrete) labelled volume using a winner-takes-all approach.

For each voxel, the program compares the fuzzy membership values across all input volumes and assigns the class label corresponding to the volume with the highest membership value at that voxel. The output is stored as byte data.

Input arguments are specified as pairs: each fuzzy volume filename is followed by its integer class label.

OPTIONS

-volume filename

Specify the output filename for the crisp labelled volume. This option is required.

-verbose

Show progress information.

-clobber

Overwrite the output file if it already exists.

-debug level

Show debugging information. Default: 0.

-help

Print summary of command-line options and exit.

-version

Print the program’s version number and exit.

EXAMPLES

Create a crisp label volume from three fuzzy classification volumes:

crispify -volume output_labels.mnc csf.mnc 1 gm.mnc 2 wm.mnc 3

AUTHOR

Vasco Kollokian

SEE ALSO

classify mincmath