User Tools

Site Tools


jvs:fscan:manual:chapter2:expression

fScan Reference Manual, Chapter 2 (CmdLineOptions): expression

Expression -- Perform arithmetic expressions on whole data sets

usage:

-expr opname varb [options]
-expr opname [options]
-expr NREPTS opname [options]

OPNAME quick reference list

Perform arithmetic or other operations on whole data sets. The primary input is the current active data set. The VARB specified in the EXPR command is the 2nd operator if the operation expects two values.

For binary operations, the second variable can be a single value, or a second data set. If it is a data set, its XY dimensions must match those of the first input data set. The Z dimension of the second input can be 1, in which case the same image will be combined with every input Z image, or it must match the Z dimension of the first input. Similarly, the T dimension of the second input can be 1, in which case the same image volume will be combined with every input time volume, or it must match the T dimension of the first input.

A summary list of all EXPRESSION operations is shown below. For more information, click on the operator name for usage details.

absvalReplace negative numbers with absolute values
adapt
+Add values to a data set
adjustNormalize local dimmest voxels
adhocCommands implemented for special (temporary) purposes
andPerform a bit-wise AND operation
bandpassApply a bandpass frequency filter
bitsCount non-zero bits in data values
boundary
calbarEmbed an intensity calibration bar into an image
censorReplace or remove selected time points in image series
clusterIdentify clusters of contiguous voxels by intensity
convolve
descend
despikeRemove intensity spikes
detransientRemove onset (offset) transient signals
detrendRemove low frequency temporal signal drift
dilateDilate around the edges of intensity regions in a 3D volume
/Divide a data set by values
eqPerform a 'equal to' data comparison
erodeErode around the edges of intensity regions in a 3D volume
expand
fftPerform a fast Fourier transform on data set
flip
floodFlood fill an intensity region starting from a seed point
freeze
gauss
gePerform a 'greater or equal' data comparison
gradient
gray
grow
gtPerform a 'greater than' data comparison
lePerform a 'less or equal' data comparison
logCalculate the log(10) of data values
ltPerform a 'less than' data comparison
map
maxFind maximum values in a data set
minFind minimum values in a data set
modulus
xMultiply a data set by values
name
neighborsCount neighbors surrounding each data point
neqPerform a 'not equal to' data comparison
normalize
notnanConvert invalid number (“nan”) values to zeros
orPerform a bit-wise AND operation
outsideIdentify values outside of an object
overlap
peak
rampGenerate image data with spatially ramped intensity values
rgb
regressRegression filter a temporal variable from a time course
roifill
roimask
roimove
rotate
scatterplot
segment
setSet part or all of data set to specified value
sine Generate image data with sinusoidal time course
smoothReduce high frequency spatial signal noise
sqrtCalculate the square root of data values
staticStore results of operation in the input data set
stretchStretch or shrink an image
-Subtract values from a data set
thin
tinterpolate
tslope
tsmoothReduce high frequency temporal signal noise
viewmaskUse a mask image to mask all slices in any 3D view
volmax
volmin
wordsize
  Op:  -  subtract operand from data
       +  add operand to data
       x  multiply data by operand
       /  divide data by operand
     min  replace data with minimum of data and operand
     max  replace data with maximum of data and operand
     gt   set data to zero if not greater than operand
     lt   set data to zero if not less than operand
     and  do bitwise AND of data and operand
     or   do bitwise OR of data and operand
    sqrt  replace data with its square root
    thin  set value in voxels around the edgle of a 3-D volume
   erode  remove (set to 0) each voxel that has at least one neighbor
          less than IN2. (IN2 must be a constant -- default=1)
  dilate  set any voxel less than IN2 to IN2, if it has at least one
          neighbor >= than IN2. (IN2 must be a constant -- default=1).
  smooth
    grow
    thin
  volmin
  volmax
normalize adjust intensities to make the local average equal IN2. IN2
          must be a constant -- default=1000).
gradient
    peak
   flood  fill a 3-D volume of contiguous voxels
boundary  find the edge of a 3-D volume of contiguous voxels
    Note: ERODE, DILATE, BOUNDARY and FLOOD, are 3-D volume operators.
    They can only be used if the input dataset (INFILE) is a single volume
    that can be completely loaded into memory. The neighbors are the 6
    adjacent voxels (or 26 -- see VAL below.) These volume operations will
    be performed on the enitre volume in the order they appear on the
    command line, before performing any of the slice by slice operations
    (e.g. -,+,x,/,min, etc.)
    Note: FLOOD and BOUNDARY assume that the volume has already been
    segmented; the low 3 bits indicate status according to:
              bit 0 (1) - ON/OFF
              bit 1 (2) - FILLED/not FILLED (or INTERIOR/not INTERIOR)
              bit 2 (4) - EDGE/not EDGE
  In2: this operand can be a file, constant, or reference variable. If it
  is a file it must have appropriate structure to match data. If it is a
  reference variable (e.g. "task2" or "motion"), the value for each image
  is used as a constant operand for that image. Note: The IN2 data set must
  have the same image dimensions as INFILE. IN2 can be bytes, short integers, 
  or floating point values. Except for constants, calculations are fastest if 
  IN2 wordsize matches the input wordsize.
  Mode: this is a character string indicating how the operand should be
  treated. It can appear right after IN2 or as the last argument. Its
  value can be:
           all       - combine each INFILE image with a IN2 image
           const     - combine each INFILE image with a constant from IN2
           meanslice - combine means of INFILE with slice indexed entry in IN2
           meancode  - combine means of INFILE with code indexed entry in IN2
           meanconst - combine means of INFILE with a constant from IN2
        For the last 3 MODE values the means are first generated based on -i,
        -b, or -K options that specify the code values for sorting the images.
        The mean images are then combined with an image in IN2. For MEANSLICE
        there should be one IN2 image for each slice, to be combined with all
        means for that slice.  For MEANCODE there should be an IN2 image for
        each sort code for each slice.
        Note: mean images are generated as floating points.
  Value: this is a modifier value that can be used when combining data sets.
  (For +,-,x, or / operands VALUE is ignored if IN2 is a constant.)
        The meaning of VALUE depends on the operand, as follows:
           + or -    -offset the result (add VALUE)
           x or /    -scale the result (multiply by VALUE)
           erode     -VALUE specifies how many neighbors need to be >= IN2
           dilate    -same as erode
           normalize -
           smooth    -whether to include 0 values
                      0 - average all voxel values (default)
                     !0 - smooth all voxels with values greater than or equal to N
           grow      -
           thin      -
           flood     -VALUE specifies how to stop filling
                      0 - fill all contiguous ON voxels, set to OFF
                      1 - fill all contiguous OFF voxels, set to ON
                      2 - fill all contiguous OFF voxels, set to FILLED
                      3 - fill all contiguous ON voxels, set to ON_FILLED
                      4 - find nearest ON voxel, then same as 0
                      5 - find nearest OFF voxel, then same as 1
                      6 - find nearest OFF voxel, then same as 2
                      7 - find nearest ON voxel, then same as 3
           boundary -same as FLOOD
  Value2: another modifier; its use is operand dependent:
           erode     -VALUE2 specifies whether to use 26 neighbors (default=0
                      unless VALUE is > 6).
           dilate    -same as erode
           normalize -VALUE2 indicates whether to include 0 voxels (default=0)
           flood or boundary - VALUE2 indicates where to start filling:
              0 - (default) explicit x,y,z value specified by "-r x y z"
                  or center if no -r specified
              1 - origin (0,0,0)
              2 - center of volume
              4 - no seed, fill entire 3D ROI (ROI specified by '-r' or
                  interactively)

See Also:

fScan Home, fScan Manual, CmdLineOptions, Manual Help

jvs/fscan/manual/chapter2/expression.txt · Last modified: 2023/02/23 18:43 (external edit)