Brain Imaging & Analysis Center

usage: resultarray = analyze(ivarb)

ANALYZE reads the input variable, IVARB, and creates an output array where the values of the array elements represent a frequency histogram of the values of elements in IVARB. The output array is of LONG type (see VARIABLES(1)). The elements in the array represent the frequency distribution of the possible values in the input. Thus, element 45 in the output array corresponds to the number of elements in the input array with value equal to 45.

By default, RESULTARRAY will be allocated 256 elements if IVARB is a byte array, or 4096 elements otherwise. If RESULTARRAY already has more than that many elements allocated to it the entire array will be used as the output histogram. In general, the length of RESULTARRAY should be 1 more than the largest integer value in IVARB.

Any IVARB values less than 0 will be counted in the 0 element of RESULTARRAY, and IVARB values greater than the length of RESULTARRY will be counted in the last element of RESULTARRAY.

The array produced by ANALYZE is useful for measuring various descriptive properties of the input variable. For example, it can be used to determine the optimal contrast enhancement to perform for a given image matrix, or as a way of computing the average value of the input.

Some image processing hardware is capable of performing an ANALYZE operation on an entire image at the hardware's video refresh rate. If the hardware is capable of this, and you run ANALYZE on an appropriate hardware image variable, then CIGAL automatically lets the hardware perform the computation. Otherwise, the software will do the work, but it may be a bit slow for very large variables.

As an example, if you have a byte matrix called VGA0, which contains a 320×184 image with pixel values covering the middle of the 0-255 intensity range, then the command:

             buf = analyze(vga0)

might produce the following array BUF:

    0    0    0    0    0    0    0   0    0    0    0    0    0    0    0
    0    0    0    0    0    0    0   0    0    0    0    1   33   65   74
   78   61   74   62   97   97   92  98  132  120  164  121  192  220  204
  288  294  389  421  461  545  673 628  815  691  971 1022  988 1086 1257
 1107 1006 1171 1321  901 1084 1042 891 1133  962  949 1025 1042 1165 1038
 1145 1084 1245  928  960 1066 1052 786  953  855  731  762  831  842  887
  734  828  777  879  761  891  687 690  631  664  560  502  534  464  487
  457  692  683  681  729  738  490 447  341  220  147  145  130   80   85
   74   40   52   18   29   11   10   9    2    2    1    0    0    0    0
    0    0    0    0    0    0    0   0    0    0    0    0    0    0    0
    0    0    0    0    0    0    0   0    0    0    0    0    0    0    0
    0    0    0    0    0    0    0   0    0    0    0    0    0    0    0
    0    0    0    0    0    0    0   0    0    0    0    0    0    0    0
    0    0    0    0    0    0    0   0    0    0    0    0    0    0    0
    0    0    0    0    0    0    0   0    0    0    0    0    0    0    0
    0    0    0    0    0    0    0   0    0    0    0    0    0    0    0
    0    0    0    0    0    0    0   0    0    0    0    0    0    0    0   0

The command STATS could then be used to describe BUF as:

        stats buf

which tells you that the intensities in VGA0 range from 26 to 130 with the greatest number of pixels (1321) at intensity 63.

The mean pixel value can be quickly calculated from the ANALYZE array as:

        a = sum(buf)            ; total number of pixels in VGA0
        buf = buf * ramp(256)   ; weight each intensity by its num. of pixels
        sum(buf)/a              ; the mean pixel intensity

The command MEAN(VGA0) would give the same result, but is sometimes slower.

See Also:
BIN(3), IPROC(1), STATS(3), VARIABLES(1)

CIGAL Home, CIGAL Manual, Functions List, Manual Help