[MINC-users] clamping minc files

[Andrew Janke]

Hrm, I'll try inline...

mincstats will make your histogram of a volume as such:

   mincstats -histogram out.hist file.mnc


a


---

#! /usr/bin/env perl
#
# Andrew Janke - rotor at cmr.uq.edu.au
# Center for Magnetic Resonance
# The University of Queensland
# http://www.cmr.uq.edu.au/~rotor
#
# Copyright Andrew Janke, The University of Queensland.
# Permission to use, copy, modify, and distribute this software and its
# documentation for any purpose and without fee is hereby granted,
# provided that the above copyright notice appear in all copies.  The
# author and the University of Queensland make no representations about the
# suitability of this software for any purpose.  It is provided "as is"
# without express or implied warranty.
#
# Fri Jun 20 17:30:06 EST 2003 - initial version

use strict;
use warnings "all";
use Getopt::Tabular;
use File::Basename;

my($Help, $Usage, $me, @opt_table, %opt);
my(@args, $infile, $outfile);

$me = basename($0);
%opt = ('verbose' => 0,
        'quiet', => 0,
        'clobber' => 0,
        'cutoff' => 0.01,
        'lowwer' => undef,
        'upper' => undef,
        'clamp' => 1,
        'out_floor' => 0,
        'out_ceil' => 100,
        );

$Help = <<HELP;
 | $me normalises a file between a max and minimum (possibly)
 |       using two histogram pct's.
 |
 | Problems or comments should be sent to: rotor\@cmr.uq.edu.au
HELP

$Usage = "\nUsage: $me [options] <in.mnc> <out.mnc>\n".
           "       $me -help to list options\n\n";

@opt_table = (
   ["-verbose", "boolean", 0, \$opt{'verbose'},
      "be verbose" ],
   ["-quiet", "boolean", 0, \$opt{'quiet'},
      "be quiet" ],
   ["-clobber", "boolean", 0, \$opt{'clobber'},
      "clobber existing files" ],
   ["-clamp", "boolean", 0, \$opt{'clamp'},
      "Force the ouput range between limits" ],
   ["-cutoff", "float", 1, \$opt{'cutoff'},
      "cutoff value to use -- explanations later..in %." ],
   ["-lower", "float", 1, \$opt{'lower'},
      "lower real value to use" ],
   ["-upper", "float", 1, \$opt{'upper'},
      "upper real value to use" ],
   ["-out_floor", "float", 1, \$opt{'out_floor'},
      "output files maximum" ],
   ["-out_ceil", "float", 1, \$opt{'out_ceil'},
      "output files minimum" ],
   );

# Check arguments
&Getopt::Tabular::SetHelp ($Help, $Usage);
&GetOptions (\@opt_table, \@ARGV) || exit 1;
die $Usage if ($#ARGV != 1);

# set up and check for file names
$infile = $ARGV[0];
die "$me: Couldn't find $infile\n" if (!-e $infile);

$outfile = $ARGV[1];
die "$me: $outfile exists!, use -clobber to overwrite\n" if
(!$opt{'clobber'} && -e $outfile);

my($l_pct, $u_pct, $range, $eqn);

# figure out upper and lowwer bounds if required
if(!defined($opt{lower})){
   $l_pct = $opt{'cutoff'};
   print STDOUT "*** getting lower range of $infile [$l_pct]\n" if !$opt{quiet};
   chomp($opt{lower} = `mincstats -quiet -pctT $l_pct $infile`);
   }

if(!defined($opt{upper})){
   $u_pct = 100 - $opt{'cutoff'};
   print STDOUT "*** getting upper range of $infile [$u_pct]\n" if !$opt{quiet};
   chomp($opt{upper} = `mincstats -quiet -pctT $u_pct $infile`);
   }

# figure out the formula
$range = $opt{upper} - $opt{lower};
if($opt{clamp}){
   $eqn = "((A[0] < $opt{lower}) ? 0.0 : (A[0] > $opt{upper}) ? 1.0 :
A[0]/$range)",
   }
else{
   $eqn = "A[0]/$range",
   }
# scale to required range
$eqn .= " * " . ($opt{out_ceil} - $opt{out_floor}) . " + $opt{out_floor};";

# do the rescaling
print STDOUT "*** scaling $infile in range
[$opt{lower}:$opt{upper}]\n" if !$opt{quiet};
print STDOUT "*** outputting in the range 
[$opt{out_floor}:$opt{out_ceil}]\n" if !$opt{quiet};
&do_cmd('minccalc', '-clobber', '-expression',
        $eqn,
        $infile, $outfile);


sub do_cmd {
   print "@_\n" if $opt{'verbose'};
   system(@_) == 0 or die;
   }



On 09/10/05, Marc Schoenwiesner <marc.schoenwiesner at mcgill.ca> wrote:
> Hi Andrew,
>
> > I use the attached perl script.
> It wasn't there. But anyway, I got the right clamping values for the structural
> scan from register. This helped - mritotal does a neat job now. (How do you get
> an intensity histogram of a volume?)
>
> The realignment of the functional volumes is still far from perfect. The
> remaining motion in z direction (x and y seem to be fine) is a good 5mm. Just
> to check, I registered the same data with AIR5 (compiled at a PC, with an
> analyze version of the data) and - that got rid of all visible movement. Does
> minctracc have problems with movement in z direction? Well, it's probably just
> me.
>
> Thanks for your help, Andrew!
>
> Cheers,
> Marc
>
> PS.: The fmr_preprocess help seems to have gotten mixed up a bit over the years:
> -blur         do not blur frames
> -align        perform motion correction this is the default, opposite  is
>               -noblur, do not perform motion correction
>


--
Andrew Janke      (a.janke at gmail.com || www.cmr.uq.edu.au/~rotor)
Canada->Montreal                                   Cell: +1 (514) 924 2012