Editing Scanco User Instructions (section)

====Interactive IPL====
*Command <code>$ ipl</code> starts IPL interactively, that is you see the result of each command as you run it. 
*While it is useful for testing, running large scripts interactively may interfere with others using the computer. 
<br>
*'''Reading an AIM file'''
:Before you can run any analysis on the AIM file, you need to read it into IPL and assign it a file handle. This is done using the command <pre>ipl> read in MJ7101.AIM</pre>
:A fraction of a second later, the file should have been read into memory and assigned the internal name "in". If we want to do some basic density calculations on it, we can use the command<pre>ipl> vox_ in 0</pre>
:*IPL, much like DCL, will automatically complete the command if it is given enough characters. 
:The "vox_" was expanded to the command "vox_scanco_param" (calculates voxel-counting parameters) and told you that the mean density of the cube is 114 mg HA/cm3 and a standard deviation of 283 mg HA/cm3. 
:The argument "in" is the internal name of the image.
:The argument "0" is the region. The command vox_scanco_param supports several numbered regions if, for example, you wanted to separately compute the density for the trabecular region and the cortical region.
<br>
*'''Segmenting an AIM file'''
:Some parameters, like average density, are computed from the [[Scanco_User_Instructions#Grayscale_Images|'''Greyscale Image''']]. Other parameters, like the trabecular number, require an image that has been segmented to [[Scanco_User_Instructions#Binary_Images|'''Binary Images''']], that is converted into "bone" and "not bone". 
:There are many ways to separate bone from non-bone, and parameters that work well for one dataset will not generally work for images of different samples, or acquired at different resolutions.
:One of the simplest segmentation methods "seg_gauss" is to apply Gaussian smoothing to the sample, to eliminate the noise present in the image, followed by a semi-arbitrary threshold that preserves the trabecular structure. 
<pre>
ipl> seg_gauss in seg

  -sigma                     [1.000000] >  
  -support                   [2] > 
  -lower_in_perm_aut_al      [300.000000] > 
  -upper_in_perm_aut_al      [100000.000000] > 
  -value_in_range            [127] > 
  -unit                      [0] > 2 
</pre>
:seg_gauss expects 8 parameters, and we have provided only 2. It will now prompt us for the remaining 6 one-by-one. 
:Leave everything at the defaults (simply hit "Enter") except the unit, which we will change to "2". 
:For young human trabecular bone from the distal tibia scanned at 20 um in the uCT 35, but then downscaled to 80 um, the above parameters give a reasonable representation of the trabecular structure. With other bones at other resolution, you may need to change the sigma, the support, and the lower threshold. 
:The unit of "2" simply means that we are working in units of mg HA/cm3.
:Now that the cube is segmented, we can run the same voxel-counting command as above and obtain more information. <pre>ipl> vox_ seg 0</pre>
:We now know that our combination of smoothing and threshold resulted in an object volume of 103 mm3 (i.e. volume of bone) and a bone volume ratio of 22.88%.
<br>
*'''Writing out a file'''
:we can write the segmented file out for future use. Accept the defaults for the remaining arguments. <pre>ipl> write seg MJ7101_SEG.AIM</pre>
:Now you can quit IPL and verify that the segmented file has been written out. <pre>ipl> quit</pre>


[[Image:Redirect.jpg|15px]] [[#top|''Return to Contents'']] or [[Main_Page|''Return to Main Page'']]