Cookbook

The Cookbook gives you recipes for common Slicer tasks.

1. Acquiring your data	2. Understanding and Importing your data into Slicer
3. Using DICOM data with Slicer	4. The print_header program
5. Viewing data in Slicer	6. Editing data in Slicer
7. Image Registration	8. Creating 3D Models with Slicer
9. Possible caveats	10. Saving the scene description to an XML file

1. Acquiring your data

This page is intended to help people who are working at the Surgical Planning Lab (SPL) to get their data into a safe place to begin working with Slicer. That data may be from the scanners at Brigham and Women's Hospital, or imported from other sites. Check the appropriate link below for your particular situation.

We suggest that other sites create a similar page describing their environment for data acquisition. Often, a major obstacle to using Slicer is finding your data, and getting it into a format that is compatible with Slicer.

• Recent MR scans done at Brigham and Women's Hospital
• Older MR scans done at Brigham and Women's Hospital
• Recent (within 3 months) CT scans done at Brigham and Women's Hospital
• Older CT scans done at Brigham and Women's Hospital
• Any MR or CT scans that you would like to acquire via FTP from a remote site
• Any MR or CT scans that you have stored on a CDROM that were acquired at a remote site

Recent MR scans done at Brigham and Women's Hospital

• MR1 - main clinical scanner
• MR2 - clinical and research scanner
• MR3 - cardiac and research scanner
• MRT - open magnet for surgical cases
• MRMAX - clinical scanner
• CNIMR1 - Center for Neuroimaging

Data from each scanner is stored in an image database dedicated to the scanner. Recent data is kept in the database, and is directly available. As newer data becomes available, older data is archived to tape and removed from the database. The data persists in the database for different lengths of time, depending on the volume of scans being done on the scanners. Generally, the clinical scanners are busier and data is available in the database for less time.
Here are the typical days of on-line storage for each scanner:

• MR1 : 7 days
• MR2 : 11 days
• MR3 : 10 days
• MRT : 120 days
• MRMAX : NOT ARCHIVED IN SPL
• CNIMR1 : 30 days

To work with this data, you need to know both the 8-digit Medical Record Number (MRN) and on which scanner your data was acquired.
Change to the appropriate directory for that scanner data, and make a copy of the data. (The directories containing the data are read-only.)
The SPL provides limited short-term storage in a file system located in directory /spl/tmp. You can use this "staging area" to sort your data and find the series of interest, before moving the data to another location. This short-term storage currently provides 100 gigabytes of storage. Data in this area will persist for 2 weeks from the time it is last accessed.
For longer term storage, ask your supervisor for some disk space. Most data sets are either 512x512 resolution or 256x256 resolution and from 50 to hundreds of images, so disk space needed is substantial: from 20 to several hundred megabytes per dataset.

1. Make a working directory for yourself in the temporary storage area:
   
   mkdir /spl/tmp/your_name
   
   
2. Change to the directory with the patient data. The naming schemes for the scanners are the same as the scanner names above. In this case:
   
   cd /d/MR1/images/idb/GINX/GENESIS/BWS/12345678
   
   
3. Copy the data to your working directory:
   
   cp -r * /spl/tmp/your_name

Older MR scans done at Brigham and Women's Hospital

Scanner	Command
MR1	where1
MR2	where2
MR3	where3
MRT	wheremrt
CNIMR1	wherecni

To find an older MR case use the following procedure:

1. Run a script to set up the above commands:
   
   source /local/bin/setup_search
2. Search for your data in the appropriate database. In this example, we will search the MR1 database for MRN 12345678:
   
   where1 12345678
   
   This will produce entries of the following form:
   
   /d/nile/backups/MR1/95111/tar25:12345678
   /d/nile/backups/MR1/96011/tar12:123456789
   
   The key field above is the one after MR1 (for example, 96011).
   This tells the date and tar file: 96 is the year 1996, 01 is the month 01 (January), and the last 1 is the tape number. You can use this information to determine if you have the correct dataset.
   
3. The next step is to make sure you have enough disk space for the data.
4. Then, send both the disk location where to store the data and the information about the data you want (for example, /d/nile/backups/MR1/96011/tar12:123456789) to the manager of the SPL data archive. He will restore the data from the tape to the disk location for you.

Recent (up to 3 months) CT scans done at Brigham and Women's Hospital

1. Get a BICS account (if you don't already have one). After you get your BICS account, you will be able to look up information in the hospital database and view images with your browser. However, to transfer the image to the SPL, you will need to contact the BICS manager and get your BICS account modified to include data transfer privileges. Also, make sure that your browser accepts cookies, or you will not be able to use your BICS account.
2. Point your browser at RadReport.
3. Click RadReport and enter your BICS access key (which you received from the BICS manager).
4. Enter the Medical Record Number (MRN) of the patient, or do a BWH search to find the information for recent cases. Most of this page is self-explanatory.
   The one important field that is not obvious is the one labeled "Stat" (Status) and it is important to know what the letters stand for:
   • S : Scheduled
   • I : In progress
   • C : Completed
   • D : Dictated
   • P : Preliminary Report available
   • F : Finalized Report
   • A : Report Addendum
   • X : Cancelled exam
   Reports that are not in F : Finalized are not likely to be available.
5. When you have located the dataset of interest, you can view any or all of the images in the series to determine if this is the data that you want. Select images and click View Image(s).
6. After viewing the images, return to the previous page to transfer the images. Click Transfer. Again, you can select any or all of the images to transfer.
7. Select Transfer compressed .gz and click Do it!. This will build a compressed tar file of your images. This may take a while to complete if you are getting a lot of images.
8. When the compression is done, click Click to download study. By default, the data will go to your home directory, so make sure to change this to the actual file destination.
9. Use the cd command to change to the directory where your data has been transferred. The file images.tar is the compressed tar file.
10. Extract the tar file with the UNIX command:
    tar xvf images.tar
11. Uncompress the tar file with the command:
    gunzip -r images
12. Remove the tar file with:
    rm images.tar

Older CT scans done at Brigham and Women's Hospital

Preferred procedure

1. Get a BICS account (if you don't already have one). You can do this through Suzanne Nagle (x2-6723) or Cheryl Mcdonald-Spriggs (x2-6266).
   After you get your BICS account, you will be able to look up information in the hospital database and view images with your browser. However, to transfer the image to the SPL you will need to have your BICS account modified to include transfer privileges. (Contact Bill Hanlon at x2-5938 or Dave Gigas at x2-6512.) Also, make sure that your browser accepts cookies, or you will not be able to use your BICS account.
2. Point your browser at RadReport.
3. Click RadReport and enter your BICS access key (which you received from the BICS manager). Click Submit.
4. Click Search PACS.
5. Enter the MRN and as much information that you know about the study that you want to restore. Click Submit Query.
6. A list of cases matching your search criteria will appear as hypertext links. To put your case in the retrieval queue, click the link. The case should be restored within 1 hour and will persist in the Radweb database for 30 days.
7. After waiting an hour for the case to be restored, follow steps 5-12 in the previous section to finish restoring the data.

Alternate procedure

1. Go to the CT reading area and ask for the chief technician (or someone who can restore a case to the PACS from the optical disk). (Unfortunately, we do not have an established person to handle this procedure. It is sometimes difficult to find a person who is capable of assisting in this regard.) You will need to know the patient name, MRN, and date of the exam. The data is automatically restored so that it is visible from the CT consoles.
2. When the case is restored to the PACS, it will show up in a list of patients available for display at the CT reading stations. From the console, you (or one of the technicians) can choose to send the data to remote destination SPL. This will send the data to the SPL.
3. Once the data is sent back to the SPL, it will be available in directory /spl/tmp/incoming.

Any MR or CT scans that you would like to acquire via FTP from a remote site

1. ftp -i to duke.bwh.harvard.edu (duke's IP address is 134.174.9.66).
2. login as anonymous, using your email address as password.
3. cd pub/incoming
4. binary (this puts you in binary file mode).
5. mput * (this assumes that you are in the directory where the remote images exist).
6. Locally, in the SPL:
   cd /d/ftp/pub/incoming
7. cp your files to a place to work on them, for example:
   cp your_images /spl/tmp/your_name
8. After you have made a copy of your data, make sure to remove your files from the ftp directory.

Any MR or CT scans that you have stored on a CDROM that were acquired at a remote site

1. Insert your CDROM into the CDROM reader.
2. On the same workstation, change to the directory where the CDROM is mounted. The mount point is /cdrom, so the command is:
   
   cd /cdrom
3. Look in this directory for your data. It may take a few minutes to mount the drive, so be patient.
4. When your data shows up, copy it to a "staging area" so that you can sort through it, for example:
   cp -r * /spl/tmp/yourname
5. Eject the CDROM with the command:
   eject
6. Take your CDROM.

top

2. Understanding and Importing your data into Slicer

1. Type the command:
   slicer
2. On the Main menu, click Data.
3. Click Add Volume. The Props panel should appear:

Slicer browser interface

1. Click DICOM (some of the letters may not be visible).
2. See Using DICOM data with Slicer.

1. Click Browse. Slicer will open the directory where you started Slicer from.
2. Double-click one of the images that you want to work on, and then click Apply.
   Slicer now tries to figure out what kind of data you have so it can properly load and orient your data. If it succeeds, Slicer will load your data and you will see your images. If it does not succeed, Slicer will display a message asking you to enter the header information manually and the following panel will appear:

   

Default Slicer header parameters

• The Field Of View (FOV) of your data, or the distance across your image, is the product of the Pixel Size and the Width in pixels of the image. In this case, the FOV would be .9375 x 256 = 240 mm.
• A voxel is a 3-dimensional pixel. The size of a voxel is the product of the Pixel Size width, the Pixel Size height, and Slice Thickness + Slice Spacing. In this case, the voxel size is .9375 * .9375 * (1.5 + 0.0) = 1.3183 cubic mm.
• The volume of a slice is the voxel size * Width in pixels * Height in pixels. In this case, the volume of a slice is 1.3183 * 256 * 256 = 86396 cubic mm, or 86.396 cubic cm (cc).
• The total volume of this dataset is the product of the volume of a slice and the total number of slices. In this case, the total volume of the dataset is 86396 * 60 slices = 5183760 cubic mm or 518.376 cubic cm (cc).

Please note that when you have finished editing your data, the segmented/edited images that Slicer writes out do not contain any header information. Slicer encodes this header information into a MRML (Medical Reality Modeling Language) file when you save your scene. This file describes the volumes that you have loaded into Slicer and how you have modified them using Slicer's tools.

In most cases, you will want to load more than one dataset. To load more data, simply repeat the procedure above. Slicer typically will display up to three data sets at one time: a foreground, a background, and a segmented dataset. When the Data tab is active, the data sets that comprise the current scene appear in the MRML File Contents (Current Scene) panel. This looks like:

Description of Scene

1. Click on the offending data.
2. Then right-click.
3. Select from the menu to either delete or cut and paste the data.

top

3. Using DICOM data with Slicer

To read DICOM files with Slicer:

1. On the Main menu, click Data.
2. Click Add Volume. The Props panel appears.

   

3. Click DICOM. (Not all the letters of DICOM may be visible.)
4. Click Select Dicom Volume. (A file browser window appears. This browser does not sort the file names alphabetically. If there are many files in a directory, it may be easier to type the path to the images than to use the browser.)

   

5. Select an image and click OK. A window appears, displaying patient information, studies, series, and files. (If there are many images in the directory, it may take a while for the information to appear.)

   

6. In this window, options include:
   • Select or deselect images to load. Selected images are shown in green and deselected images are shown in red.
   • Click Increasing or Decreasing to load images in increasing or decreasing order.
   • Click Preview to display images as small icon-size images in the bottom panel. You can change the size of the previewed images. The default size is 32x32 pixels.
   • Click List Headers to display header information about the geometry of each image in the series.
   • Click Check to display information about the volume and how to group the data. With DICOM data, the organization of the data is not always apparent. Check examines the slice location and thickness of each image in the volume to make sure all slices are the same distance apart. This helps to show when slices are missing. This is also useful when several DICOM acquisitions are lumped together.
7. After the data is organized properly, click OK.
8. Click Extract Header. This displays the information extracted from the headers describing the dataset.

top

4. The print_header program

• GE Signa (This is usually old data.)
• Siemens Somatom
• GE Genesis (Slicer was originally designed for this type of data, but has been expanded to accommodate other formats.)
• DICOM (Slicer can usually read DICOM (digital imaging and communications in medicine) data directly. However, since DICOM is such a complicated specification, not all DICOM data will be readable in Slicer.)
• print_header can read data in compressed (.gz or .Z) format as well as uncompressed.

print_header I.001

print_header V1.0 ENTRY AND POSITION ARE 1 1
input_prefix = I
output_prefix =
suffix = <none>
filename_format = %s.%03d
patient_name = mrsmith
patient_sex = F
date = 8/25/93
time_hr_min_sec = 07:01:32
study_desc = *SPL*LINOMIDE
series_desc = *SPL*FILTER(2d2chan)
hospital_name = BRIGHAM WOMENS HARVARD BAY 2
patient_id = 12345678
exam_modality = MR
image_type_text = genesis
image_type = N/A
sop_instance_uid = N/A
image_time = N/A
manufacturer = N/A
referring_physician =
physician_reading_study = N/A
manufacturers_model_name = Ü
anatomic_structure = N/A
birth_date = N/A
weight = N/A
body_part_examined = N/A
repetition_time = N/A
echo_time = N/A
inversion_time = N/A
reconstruction_diameter = N/A
generator_power = N/A
series_in_study = N/A
images_in_acquisition = N/A
rescale_intercept = N/A
rescale_slope = N/A
requested_proc_desc = N/A
patient_age = 36
x_resolution = 256
y_resolution = 256
bytes_per_slice = 131072
header_size = 7900
slice_number = 1
image_type_num = 2
byte_order = 0
status = 1
number_echoes = 2
echo_number = 1
compressed = 0
first_slice = 1
last_slice = 1
num_missing = 0
bytes_per_pixel = 2
num_bytes_data = 131072
num_bytes_header = 7900
bits_stored = 65535
high_bit = 65535
bits_allocated = 65535
patient_position = HFS
patient_orientation = N/A
number_of_slices = 1
exam_number = 56168
series_number = 1
gantry_tilt = 0.100000
pixel_xsize = 0.937500
pixel_ysize = 0.937500
fov = 240.000000
aspect = 3.200000
thick = 3.000000
space = 0.000000
image_location = -20.000000
coord_center_r = 0.000000
coord_center_a = 0.000000
coord_center_s = -65.000000
coord_normal_r = 0.000000
coord_normal_a = 0.000000
coord_normal_s = 1.000000
coord_r_top_left = 120.000000
coord_a_top_left = 120.000000
coord_s_top_left = -65.000000
coord_r_top_right = -120.000000
coord_a_top_right = 120.000000
coord_s_top_right = -65.000000
coord_r_bottom_right = -120.000000
coord_a_bottom_right = -120.000000
coord_s_bottom_right = -65.000000
imagepositionpatient = 0.000000 0.000000 0.000000
imageorientationpatient = 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000


• The information may not be correct, and will be inaccurate if you change the images without changing the headers.
• Not all of the parameters have meaning for all datatypes.

top

5. Viewing data in Slicer

• A single series of grayscale images
• A complex scene comprised of several grayscale and segmented data sets
• 3D models, such as the composite scene below of a grayscale MR dataset of an orange, with a single section of the orange (in red) and the rest of the orange (in orange)

The top portion is the 3D Viewer.
The bottom portion is the 2D Viewer.

Slicer has many controls over the 3D Viewer and 2D Viewer. Here are some of them.

Selecting the View mode

1. Click View. A drop-down menu of available view modes appears.
2. Select a view mode from the menu. The view will change to reflect your selection.

Controlling the 3D view

• Left-click the view and move the mouse to rotate the view.
• Click the middle button (if available) on the view and move the mouse to move the view left, right, up, or down.
• Right-click the view and move the mouse to change the size of the view.

Annotations on the scene

To control the annotations over images:

1. Click More. A drop-down menu of modules appears.
   

2. Click Anno. A panel appears.
3. Select the annotation options you prefer.
4. To save options, click File, then click Save Current Options. This will save your options to a file called Options.xml.

Controlling the 2D view

The 2D view of the scene (see below) can display multiple data sets in a variety of ways. Each of the three separate windows displays the data as a view from one of the three major axes.
Note: Changing the 2D view may also change the 3D view.
Here are some of the ways to control the 2D view:

• To control whether a slice is visible in the 3D view, click V for that window. This is useful to view the orientation of multiple data sets and models with respect to each other.
• To select a slice in a volume, move the slider for that window.
• You can view up to three data sets at one time. To control how each dataset appears:
  • Click Fg to select the dataset in the foreground.
  • Click Bg to select the dataset in the background.
  • Click Lb to display a label map layer.
  • Click Or to change the default orientation of the data.
• To change the window/level for one of the views, right-click the view. A window appears. Adjust the sliders to change the display.

• In the first window, slice 0 is selected; the view is visible in the 3D view; the orientation is Axial; the grayscale data is in the background; the skin is in the foreground; and the label of one section of the orange is loaded as a label map (blue outline).
• In the second window, slice -15 is selected; the view is not visible in the 3D view; the orientation is Saggital; the skin is in the background; the section of the orange is in the foreground; and there is no label map.
• In the third window, slice -3 is selected; the view is not visible in the 3D view; the orientation is Coronal; the grayscale data is in the background; the section of the orange is in the foreground; and there is no label map.

Controlling the scene with the user interface panel

Move the cursor over the Slicer icon:
The user interface panel appears:

• The slider on the left between Fade and Exit is useful when you have data in the foreground and background. When the slider is at the top, it will display only the foreground dataset; when it is at the bottom, it will display only the background dataset. When the slider is in between, it will blend the two data sets. This is particularly useful when you are registering two data sets, and want to see how accurate the registration is.
• The head icon rotates the 3D view. Click R (Right), L (Left), P (Posterior), A (Anterior), I (Inferior), or S (Superior) to rotate the image in each direction.
• The arrows pointing up, down, left, and right can rotate the 3D view in the appropriate direction. Click the arrow and the 3D view will rotate.
• FOV stands for Field Of View, which is the size of your image. If the image is too small, enter a larger value for FOV. If the image is too large, enter a smaller value for FOV.
• Save 3D saves the current 3D scene. Enter a name for the scene in the field. Slicer saves the scene to the directory where you started Slicer.
• Parallel toggles between a parallel projection (which preserves parallel lines, and displays an object the same size regardless of how far from the observer) and a perspective projection (which displays closer objects as larger than farther objects).

top

6. Editing data in Slicer

Slicer has a set of effects (tools) for editing (or segmenting) your data.

Before Editing

• If you have multiple copies of grayscale or label maps that are part of your scene, select the one you currently want to edit.
• Enter a Descriptive Name that is meaningful for the data (such as 'vessels' or 'ventricles').
• Click View to change the view mode to 4x512.

NOTE: Slicer will strip the headers off the images when you save your data. Slicer encodes the geometry and orientation of your data in a MRML (Medical Reality Modeling Language) file. You can save, reload, and repeatedly update your MRML file until you have created the scene you want.

Click Start Editing. The Effects tab appears:

Editing Data

About the Editing Effects

• Select the correct input volume of data to edit, so that the edits will be applied properly.
• Set the scope of your edits properly: one slice (the one on the screen that you are currently editing); three slices (the current slice and the slices on either side of the current slice); or the entire volume. For some effects, such as drawing, it only makes sense to have the scope of the edits be the current slice. For other effects, such as thresholding, it makes sense to apply the changes to the entire 3D volume.
• When you select any of the effects, the following row of buttons appears: The button labels are abbreviations for the effects (except for PhaseWire and Measure Island). The selected button is the current effect. A description appears under the row of buttons.
  The right-most button is Undo (Un), which becomes available after you apply any effect. If you click Un, it undoes (or reverses) the last applied effect.

Using the Effects

• Threshold
• Change Label
• Change Island
• Measure Island
• Erode
• Draw
• Save Island
• PhaseWire
• Remove Islands
• Identify Islands

Threshold effect

1. Click Threshold. This window appears:
2. Select the Input Volume.
3. Select the Scope and Interact ranges.
4. Click Output and select a color, or enter an Output color number. You can also enter an optional label for what that color will represent. For instance, in this example, the chosen color is 2 (orange) and the label is 'skin'.
5. Adjust the Lo slider to exclude points whose signal intensity is too low.
6. Adjust the Hi slider to exclude points whose signal intensity is too high.
7. Click Apply to apply the thresholding effect. Slicer will update the Working volume with your new label of thresholded data.

Results of thresholding

Change Label effect

1. Click Change Label. This window appears:
2. Select the Input Volume.
3. Select the Scope range.
4. Enter a Value to change. This is the value of label of pixels in the Input Volume that Slicer will change.
5. Click Output and select a color, or enter an Output color number. This is the new value that pixels with the Value to change will receive.
6. Click Apply to apply the change label effect. Slicer will update the Working volume with your newly labeled data.

Identify Islands effect

Change Island effect

1. Click Change Island. This window appears:
2. Select the Input Volume.
3. Select the Scope and Render ranges.
4. Click New Label and select a color, or enter a New Label color number. This is the new label value that pixels in the selected island will receive.
5. Click any part of the island. The pixels that are part of that island will receive the New Label value.

Results after using Change Island:

Measure Island effect

1. Click Measure Island. This window appears:
2. Select the Input Volume.
3. Select the Scope range.
4. Click Island Label and select a color, or enter an Island Label color number. Measure Island will calculate the volume of the pixels with the Island Label color in the selected island.
5. Click any part of the island. Slicer will display the Size, in pixels, of the island you clicked. Slicer will also display the size of the Largest island within the selected Scope.

To calculate the total volume of the selected island, use the following formula: size * pixel_size * pixel_size * (thickness + spacing)

Erode effect

1. Click Erode. This window appears:
2. Select the Input Volume.
3. Select the Scope range. Multi Slice applies the one-slice erosion to each slice. 3D treats the whole volume at once.
4. Click Value to Erode and select a color, or enter a Value to Erode color number. Erode will act only on pixels with the selected Value to Erode color.
5. Enter 0 for the Fill value.
6. Enter the number of times to apply the effect in the Iterations field.
7. Select 4 for Neighborhood Size for less erosion or 8 for more erosion.
8. Select Active to apply Erode to the active view. Select Native to apply Erode to the original view.
9. Click either:
   • Erode to remove pixels with the Value to Erode from around islands.
   • Dilate to add pixels with the Value to Erode around islands.
   • Erode & Dilate to erode, then dilate.
   • Dilate & Erode to dilate, then erode.

Results after using Erode:

Draw effect

1. Click Draw. This window appears:
2. Click Output and select a color, or enter an Output color number. This is the new value that pixels you draw on will receive.
3. Enter a value for the Point Radius, which controls how fine each line of the drawing is. Using a value of 1 or 2 allows you to select individual points.
4. Select the Mode of the left mouse button in the draw window. Draw is for drawing on points. Select is for selecting points. Move is for moving selected points.
5. If the Mode is Select or Move, you can use:
   • Select All to select all points.
   • Deselect All to deselect all points.
   • Delete Selected to delete selected points.
   • Delete All to delete all points.
6. Select the Render range.
7. Select the Shape that the left mouse represents in the draw window.
8. Draw on the region of interest (ROI), using the left mouse button. You can use the right mouse button to zoom in and out on the area. The middle mouse button moves the image in the window.
9. To clear the list of points after you click Apply, click Yes. To keep the same points in a list after you click Apply, click No.
10. When you have finished drawing the outline of the ROI, click Apply. Slicer will fill the ROI with the Output color.

Remove Islands effect

Save Island effect

PhaseWire effect

1. Click PhaseWire. This window appears:
2. Click Basic. (Only the Basic features work.)
3. Click Output and select a color, or enter an Output color number. This is the new value that pixels you outline will receive.
4. Select the Render range.
5. Select the Shape of the outline. Polygon will fill in the feature when you click Apply. Lines will create only the outline and not fill in the feature.
6. Select the Image Feature Size.
7. Click the left mouse on the edge of the feature that you want to outline. Release the mouse button, but follow the edge with the mouse.
   If the "wire" starts to stray from the edge, click the mouse button to pull in the wire.
8. To undo the last mouse click, click Undo Last Click.
9. To start the outline process again, click Clear Contour.
10. When you have finished drawing the outline of the ROI, click Apply. If Polygon is selected, Slicer will fill the ROI with the Output color. If Lines is selected, Slicer will outline the ROI with the Output color.

top

7. Image Registration

1. Click Data on the Main menu. The Data panel appears.
2. Select the volume(s) and/or model(s) that you would like to move with this transform.
3. Click Data.
4. Click one of the volumes. This will highlight the volume.
5. Click Add Transform to surround them with a transform. This will create an identity matrix.
6. Click Alignments.
7. Click Manual.
8. Move the slider in the user interface panel at the bottom-left (near the Exit and Fade buttons) so that both volumes are visible.
9. Use the translation and rotation sliders to move one dataset into alignment with the other. You can also drag one dataset to align it with the other dataset.

Example of Registration

If we save this scene and examine the resulting .xml file, there is a text entry similar to the following:

Matrix name='manual7' matrix='1 0 0 -9 0 1 0 -12 0 0 1 105 0 0 0 1'

Automatic Registration

1. Click Alignments.
2. Click Auto.
3. Select one of the matrices generated in the manual registration phase.
4. Select the reference volume and the volume to move.
5. Click Run. The automatic registration will run. This will take a few minutes, including initialization and a series of displays that appear and disappear.

If we save this new scene and examine the resulting .xml file, there is a new text entry for the matrix:

Matrix name='manual7' matrix='0.999995 0.001058 0.003264 -8.39007 -0.001059 1 0.000422 -11.1639 -0.003263 -0.000425 0.999996 102.155 0 0 0 1'

top

8. Creating 3D Models with Slicer

1. From the Main menu, click ModelMaker. The ModelMaker panel appears.
2. Click the Create tab. The Create panel appears.

3. Select the Volume.
4. Click Label and select a color, or enter a Label color number. This is the value that will be used to create the model. This should correspond to the value in the label map.
5. Enter a Name for this model.
6. Enter a Smooth value, which controls how smooth the resulting model will be.
7. Enter a Decimate value, which indicates how redundant polygons will be reduced.
8. Click Create to start the model creation process.

Below is the resulting 3D skull model.

top

9. Possible caveats

Some Slicer caveats and some possible solutions

Slicer is a research effort and a work in progress. As such, there are many features that don't always work in all cases. Here are some known issues and some ways to deal with them.

• My data is 512x512 resolution and this seems to confuse Slicer.

  Originally, Slicer was designed for images that are 256x256 resolution. However, Slicer has been modified to accommodate 512x512 and larger images. If your data is 512x512, and Slicer is still having difficulties handling it, try reducing the resolution of your data to 256x256 resolution. If you do this, please realize that:

  • Your reduced number of pixels will now be twice as big as they were, since there are half as many pixels in the X direction and half as many in the Y-direction after reduction.
  • In most cases, the reduced data sets look visually acceptable. But, since there is one-fourth the information in each reduced slice, some smaller structures may not be acceptable for viewing or editing.

• The headers on my data are confusing Slicer.

  Some users find that things work best if they record the information about their images manually, and then throw the headers away. Any data that you create and save with Slicer will be written out in no-header format and numbered slice.001 to slice.xxx where xxx is the highest numbered slice.

• When I try to save my changes, Slicer crashes and prints a "Segmentation fault" message.

  Segmentation fault is usually caused by one of two reasons:

  • The user is trying to write data to a directory that is owned by another user.
  • The user is trying to write data to a full file system. To check if the file system is full, change to the directory where the data is and try to copy one image to a temporary file in the same directory (for example, cp I.001 junk). If this does not work, then the file system is likely full. Note that the df command does not always give the correct results, so the above approach is preferred

top

10. Saving the scene description to an XML file

• Save Scene
  This command will save the current session with a default name data.xml. If the user has previously specified a name, it will save the file as name.xml.
• Save Scene As
  Allows the user to specify a name for the .xml file. Once the user has specified a name, use Save Scene for subsequent saves.
• Save Current Options
  Allows the user to save a file titled Options.xml that contains the options selected during the current Slicer session (for example, the FOV, window settings, and so forth).
• Save Scene with Options
  Allows the user to save the scene (see Save Scene or Save Scene As above) and the options associated with the current Slicer session.

How to use these save options:

1. Launch Slicer from the directory where you would like to save the label maps, .xml files, and models.
2. Load the data that you want to segment, and select values for viewing, such as FOV.
3. Use Save Scene As to save an .xml file with a specific name.
4. Use Save Scene with Options to save the scene with your selected options.
5. Begin segmentation. Save the label maps.
6. Use Save Scene with Options to save the scene.
7. Any time that you change and save the label maps, use Save Scene with Options to save the scene.

top