Introduction to the 3D Slicer platform

April 19th, 2016

Alexis Girault, R&D Intern in Medical Computing

What will be covered
● Introduction to 3D Slicer
● Loading Data
● Nodes & Module architecture
● Layout and Slice Viewers
● Volume visualization
● Modules in pipeline
● Saving/Closing a scene
Credits
This slides integrate material available at:
www.slicer.org/slicerWiki/index.php/Documentation/4.5/Training

It is built on prior work of:

● Jean-Christophe Fillion-Robin, Technical Expert, Kitware Inc
● Julien Finet, R&D Engineer, Kitware Inc
● Sonia Pujol, Ph.D., SPL/BWH
● Dominik Meier, Ph.D., SPL/BWH
● Ron Kikinis, M.D., SPL/BWH

Portions of this work have been obtained under 3D Slicer Contribution and Software License Agreement
PART 1

Introduction to 3D Slicer
What is 3D Slicer
● Software platform for:
○ analysis (registration, interactive segmentation)
○ visualization of medical images
○ research in image guided therapy

● Extensible: powerful plugin capabilities

● Free + Open Source + Cross Platform

Behind the scene

Slicer is built every night on:

● Windows
● Linux
● MacOSX
Online documentation

www.slicer.org/slicerWiki/index.php/Documentation/4.5
Over 100 modules

Slicer 4.5 contains over

100 modules for image
segmentation, registration, and
3D visualization of medical
imaging data.
PART 2

Nodes & Modules architecture

Data Types (1/3)
Volume: structured grid, 3D image based of
voxels of a certain size (image spacing)
● Scalar volume: continuous voxel values
● Label Map: discrete integer voxel values
● Formats: mha, mhd, nrrd, dicom, ...

Model: polygonal mesh, 3D surface

● 3D points and triangles
● Points can hold data
● Formats: vtk, vtp, stl
 ar
Data Types (2/3) cess
ible
in th e t oolb

Ac

Annotations: interactive markers

● ROI : Region of interest (useful for cropping)
● Fiducial: defines a 3D position (markups)
● Ruler: measure distance between two positions
Data Types (3/3)
And much more:
● Transforms
● Lookup tables
● Transfer Functions
● Tables
● Double Arrays
● ...

Not really used in NIRFAST pipeline

MRML Nodes & Scene
● MRML Library provides API for managing medical
image data types (Volumes, Models, Transforms,
Fiducials, Cameras, etc) and their visualization.

● Each data type is represented by a special MRML

node.

● A MRML Scene is a collection of MRML nodes.

Modules (1/3)
● 3D Slicer plugin architecture allows to easily extend the
core platform. Those plugins are called modules.

● Those modules can be implemented in:

○ C++
○ Python
○ Matlab (link to NIRFAST-Matlab)
Modules (2/3)
Choose your modules from
the module drop-down menu:
● Magnifying glass to search a module by name

● A list of all modules can be accessed at the top

● Main modules are in the section right below

● All modules can be accessed by category

Modules (3/3)
How it works:
● Set input nodes (ex: CT volume)

● Set input values (ex: threshold value, options)

● Set output nodes (ex: new empty node)

● Apply

➔ Output nodes updated based on inputs

Sometimes: user interaction needed (ex: segmentation)

PART 3

Loading Data
Loading Data (1/2)
Two options :
● File > Add Data
● Drag & Drop file in the application window

Same result : creation of a new object node

 Loading Data (2/2)

This popup appears.

Click here to add data
Loading DICOM (1/3)
Two options :
● Module DICOM > Show DICOM Browser > Import
● Drag & Drop folder containing DICOM images

Same result : imports patient info in DICOM Browser

 Loading DICOM (2/3)
Press Add Link to import
the directory

Drap & drop DICOM Press OK to load directory Press OK to finish import
folder inside Slicer into DICOM database
Loading DICOM (3/3)

Select the dataset to load in

the DICOM Browser

Press Load to load it in the

scene as a MRML node
Data module
List of nodes, allows you to:
● Edit node properties
redirects you to the appropriate module
● Rename node
● Delete node
● Inspect node (bottom)
PART 4

Layout and Slice Viewers

 User Interface Overview
Main menu

Toolbar
3D Viewer
Module GUI

2D anatomical
slice viewers
Data Probe
 Layout Manager

Press the
layout icon in
the toolbar to
select a layout
Link Slice viewers (1/2)
Click on the link icon Click on the eye icon
to link all three 2D to display the slices in
Viewers. the 3D Viewer.
Link Slice viewers (2/2)
Slice viewer toolbar (1/3)

Position your mouse

over the pin icon to display
the slice viewer toolbar.

Re-center your slice within

the window.
(useful after changing the
displayed volume)
Slice viewer toolbar (2/3)
This shows the background
volume and the view
names.

Press the >> button to

open the full viewer toolbar.
Slice viewer toolbar (3/3)
The viewer bar has:
● 3 lines for:

○ Label Map

○ Foreground

○ Background

● 3 columns for:
display borders of labelmap only
○ Opacity

○ Interpolation (on/off)

○ Name of displayed node

Select displayed slice
Move the slider to display slice along the viewer axis.
Display multiple volumes
Click on this icon to display multiple slices at once.
View controllers module

You can also access the

viewer toolbars in the view
controllers module.
PART 5

Volume visualization
 Move volume in Viewer
You can displace a slice in its view by pressing a specified key and dragging
your mouse like explained below:

Reset to original: Zoom in: Zoom out: Translate:

Press Shift + drag up Shift + drag down Ctrl + drag mouse
 Window level
You can update the window level by dragging the mouse in the 2D Viewers.
(To display the scalar bar, go in module Data Probe > Scalar Bar > Enable)

Darker: Brighter: More contrast: Less contrast:

Drag mouse up Drag mouse down Drag mouse left Drag mouse right
Volumes info/display
Volumes module offers:
● Volume Information
○ Image dimensions, spacing, origin
○ Scalar Range
○ ...

● Display parameters
○ Select lookup table (scalar value to RGB)
○ Interpolate voxel values (smoothing)
○ Select Window Level (manual, automatic)
○ Set Threshold
Volume Rendering (1/2)
Select the Volume to render in
3D and click on the eye icon

You can select a Preset, slide

the Shift slider, or unselect
Shade to get a better rendering

Click on Synchronize with

Volumes module to use the
window level of the 2D viewer

More advanced options if

needed
Volume Rendering (2/2)
Use those options to
manipulate the Region
Of Interest that was
created with the volume
rendering.

You can manipulate the

ROI annotation in the 3D
and the 2D Viewers.

Uncheck Shade to
properly visualize within
the volume.
PART 6

Modules in pipeline
Crop volume
● Role: crop volume to region of interest

● Input volume: select volume to crop

● Input ROI: either:

○ Create new AnnotationROI
○ created in volume rendering
○ created with annotation

● Crop! Creates a cropped file named:

“inputname”-scale_1
Cast Scalar Volume
● Role: convert the volume voxel scalar
values to short type (required for
automated segmentation techniques)

● Input volume: cropped volume

● Output volume: either:

○ Create new Volume (as…)
○ same as input (overwrites it)

● Apply: updates the output volume

Editor - Segment Tissue (1/3)
When selecting the Editor
module, a pop-up shows up to ask
you what color table to use. Select
GenericColors and press OK

Make sure the Master Volume is

the volume you wish to segment,
if not, select the good node here

This will create a node for your

label map in Merge Volume
Editor - Segment Tissue (2/3)
● Tools: select your tool (drawing
tools, filters…)

● Undo/Redo: allows you to go back

step by step if you are not satisfied
with your changes

● Label: active label for the tool

● Options: options of the tool/filter

● Apply: apply the filter (filter only)

● ?: information on the selected tool

Editor - Segment Tissue (3/3)

For detailed documentation on each tool and filters, go to:

www.slicer.org/slicerWiki/index.php/Documentation/4.5/Modules/Editor

Web search keywords:

Slicer wiki editor module
Create Mesh
● Role: create NIRFAST volumetric mesh
and allows you to place the light sources
and detectors

● Input Label Map: segmented volume

● Sources/Detectors: placed fiducials

● Output Mesh Dir/Name: path where

the volumetric mesh will be saved

● Apply: launches NIRFAST-Matlab

window
Import Optical Properties
● Role: extract the point data of the VTK
volumetric mesh and resamples it in
volumes for each optical property

● VTK Mesh: VTK file created after light

reconstruction in NIRFAST-Matlab

● Bounding volume: either:

○ original cropped volume
○ none (will need to specify spacing)

● Apply: dynamically outputs volume

nodes and display them accordingly
PART 7

Saving/Closing a scene
Saving a scene (1/3)
Saving a scene (2/3)
Saving a scene (3/3)
Closing current Scene
Open source scientific computing
Collaborative Software R&D
● Algorithms & applications
● Software process & infrastructure
● Support & training
● Open source leadership
Supporting all sectors
● Industry, government & academia
 Customers & Collaborators
Over 75 academic institutions Over 50 government agencies Over 100 commercial
including: and labs including: companies in fields including:
Automotive
Harvard National Institutes of Health (NIH) Aircraft
MIT National Science Foundation (NSF) Defense
University of California, Berkeley National Library of Medicine (NLM) Energy technology
Stanford University Department of Energy (DOE) Environmental sciences
California Institute of Technology Defense Advanced Research Finance
Imperial College London Projects Agency (DARPA) Industrial inspection
Johns Hopkins University Army Research Lab (ARL) Oil & gas
Cornell University Air Force Research Lab (AFRL) Pharmaceuticals
Columbia University Sandia (SNL) Publishing
Dartmouth College
Robarts Research Institute
University of Pennsylvania
Rensselaer Polytechnic Institute
University of Utah
University of North Carolina
Successful small business
● Founded in 1998, privately owned
● More that 110 current employees, with 37% MSc, 47% Ph.D
● $21.3 M revenue in 2013

Offices
● Clifton Park, NY
● Carrboro, NC
● Santa Fe, NM
● Lyon, France
Awards
● 2016 Winner of the Top Workplaces Award

● 2010, 2011, 2012, 2015 HPCwire Best HPC

Visualization Product or Technology awarded to
VTK and Paraview

● 2007 to 2013 Inc. Magazine's list of 5,000 fastest-

growing private companies

● 2011 Inc. Top Small Company Workplaces Award

● 2011 Tibbetts Award for Outstanding Research

Current Openings
We're always looking for talented and
motivated people. For more information
about Kitware and available positions,
visit http://jobs.kitware.com