Slicer3 and Slicer4

1
Slicer3 and Slicer4

• Steve Pieper, PhD
• Chief Architect for 3D Slicer

Engineering Core PI for Neuroimage Analysis
Center Engineering Site PI and Dissemination
Co-PI for National Alliance for Medical Image
Computing CEO of Isomics, Inc.
2
Topics

• Overview of 3D Slicer
• Slicer3 to Slicer4 Transition Issues
• Radiation Therapy Implementation Issues
  (Discussion Topics)

3
Core 1b Engineering5 Aims / 5 Platforms
Architecture tools, operating paradigms,
reporting mechanisms, integration points End-user
platform interactive methods and information
visualization for longitudinal analysis,
exploratory data analysis, and translational
research Computational platform stream
processing, cloud computing, statistical
analysis, informatics, machine learning Data
management non-imaging and derived data, DICOM
and cloud services Software engineering and
software quality navigable timeline for
revision control, build, test, documentation and
release
4
What is Slicer?

• Multi-OS End-User Application (Windows, Linux,
  Mac)
• Medical Image Visualization and Analysis
• Multi-Modal CT, MR, fMRI, dMRI
• Integrated View Images, Surfaces, Annotations,
  Devices
• Project on the Scale of a Microsoft Office or
  Adobe Photoshop
• Extensible Architecture
• Dozens of Custom Modules
• Application Specific Functionality
• Fully Open and Non-Restrictive License
• All Source Code Available
• Can Be Used in Commercial or Proprietary Projects
• Common Platform for Medical Image Computing
  Research and Dissemination
• Because Slicer Exists, Researchers Dont Need to
  Reinvent the Wheel for Each Project

http//slicer.org
Images A. Jakab and the Slicer Community
5
End-user platform

• 3D Slicer a cross platform system for
  translating innovative algorithms into clinical
  research applications

• Developer Needs
• Clinical Grade IO and UI
• Rich Utility Libraries
• Extensible and Reconfigurable
• Stable Base
• Easily Deployable

User Needs Easy Install and Upgrade Standard
Clinical Behavior Consistent Interface Advanced
Functionality
6
Developer Needs

• Clinical Grade IO and UI
• Slicer developers work closely with DBPs on real
  data and are exposed to clinical state of the art
• Rich Utility Libraries
• NA-MIC Kit
• Extensible and Reconfigurable
• Loadable Extensions
• Scriptable GUI
• Stable Base
• Nightly Extensions Built Against Latest Slicer
  Release
• Slicer Release Relies on Upstream Release (ITK,
  VTK, etc)
• Easily Deployable
• Binary Installers for the most common platforms

7
User Needs

• Easy Install and Upgrade
• Binary installers and Extension Wizard
• Standard Clinical Behavior
• Adhere to Conventions where Possible
• Varies by Domain (IGT Navigation vs. Neuroscience
  Research vs. RT Planning)
• New Functionality has no Standard by Definition
• Consistent Interface
• Learning one Slicer Module should help you know
  how to use other ones
• Advanced Functionality
• Main Motivation you use slicer because nothing
  else does what you need we try to make that as
  painless as possible

8
Translational Focus

• Improved DICOM Support
• Local Database
• Query / Retrieve / Store
• Application Hosting
• OpenIGT Link and Real-Time Performance
• Clinical Interfaces
• Uncluttered Display
• Compare View, Lightbox
• Large Format, Multi-Monitor

9
Slicer Development Dissemination

• Distributed Team
• Twice-Yearly Project Weeks with 100 Developers
• Yearly Large Releases Integrate Latest Advances
• Slicer 3.6.3 Released March, 2011
• Training and Outreach Events
• Over 30 in 2010 Alone!
• Slicer Extension Modules
• Standard Framework for Disseminating our Science
• End-to-End Solutions Contributed by Dozens of
  Institutions
• Publications, Code, Data, Tutorials
• Reproducible Science

10
Slicer Impact

• 20,000 Downloads Per Year
• 200 Peer Reviewed Publications
• Dozens of Funded Collaborations
• Active Academic and Industrial Participation

Software is Increasingly Essential for Medical
Research Slicer has a Proven Formula for
Software Innovation and Dissemination
11
Slicer4 Collaborations

• Nokia - Qt Software Ecosystem
• CTK (Common Toolkit)
• Widgets, DICOM, Messaging
• BRAINS, BioImageSuite
• ITKv4
• Scipy (Scientific Python), Nipy (Neuroimaging
  Python)
• And, of course, NAC, NCIGT, NA-MIC DBPs,
  Collaboration Grants, Like-Minded Developers,
  Longtime Friends and Family

12
Slicer4 Compatibility Pledge

• Import .xml and .mrml from Slicer2.x and
  Slicer3.x
• Command Line Modules will not Need to Change
• Except perhaps the build tree
• Can be modified for new features if desired
• Core Modules will be Ported
• Help will be Provided for Porting Loadable
  Modules
• MRML and Logic Basically Unchanged, Only GUI
  Needs to be Rewritten
• ModuleWizard to Create Code Framework
• Consistent Slicer Terminology and Concepts
• Ease User Transition, Leverage Training

13
Slicer4 Git

• Git Advantages
• Arguably better than svn
• Many (all?) of our colleagues are migrating
• Git Disadvantages
• Best practices still evolving
• Many (most?) of our colleagues are frustrated )
• We had planned to use git from the start for
  slicer4, but have backed off.
• Now we suggest using git-svn for your own work
  and collaborative topic branches
• http//www.slicer.org/slicerWiki/index.php/Slicer3
  git-svn

14
Slicer4 Qt

• Qt for Slicer4
• Released LGPL by Nokia
• Large Developer Community
• Many Highly Refined Tools
• Major Effort Underway thanks to ARRA Supplement
• Results are Very Encouraging, but there is a lot
  To Do
• Qt as a Platform
• No Longer in Nokias Critical Path for
  Smartphones
• Desktop Qt Works Well for Our Needs
• Need to Watch Long Term Development Trends

15
Slicer4 Qt Impact

• No More X11 on Mac OSX
• New C Paradigms
• Designer for GUI Construction
• Native WebKit
• Native HTML (CSS, HTML5, webgl)
• JavaScript Engine

16
Slicer4 CTK

• See commontk.org for info
• CTK is Home Many for Generic Parts of Slicer4
• CTKWidgets, including Transfer Function Editor
• ctkDICOM, based on DCMTK
• Other Code may Migrate (Execution Model? MRML? )
• Active International Development Team

17
Slicer4 PythonQt

• Developed by Mevislab (Germany)
• LGPL, Stable
• Scripted Access to Classes, Signals, Slots,
  Properties
• May (or may not) be Replaced with PySide in the
  Future

18
Slicer4 Tcl and Python

• VTK and VTK-based Slicer4 Code are Wrapped with
  Python
• A Compatibility Layer Allows Tcl to Call
  VTK/Slicer Code via Python
• This is the inverse of Slicer3, where Python
  calls are routed through Tcl
• Used for Editor and Slice Interaction
• This layer is for porting only and should not be
  used when writing new code
• Any Tk or KWWidgets GUI code will need to be
  rewritten in Qt

19
Slicer4 SuperBuild

• Evolution of getbuildtest.tcl Concept into Pure
  CMake
• More General Implementation to Serve as Template
  for Other Projects
• getbuildtest is Still the Supported Tool for
  Slicer3.6
• Hierarchical
• CTK also Uses SuperBuild
• Extension Support
• Coming
• Try It
• Slicer4 Build Instructions on Wiki

20
Slicer4 Development

• Emphasis on the NA-MIC Software Engineering
  Methdology
• Testing, Coverage, Documentation Coding Style
• New Modules Start as Extensions
• Only Widely Applicable Functionality Bundled with
  Application
• Only When Coding Standards Are Met
• Increase Communication via the Developer Mailing
  List
• Keep Everyone in the Loop on Architecture and
  Coding Conventions
• Verbose Commit Messages

21
Slicer4 What Works (mostly)

• Main GUI, Toolbars, Menus
• Scene/Data Load/Save
• Slice and 3D Visualization (pan/zoom, overlays,
  models)
• Command Line Modules
• Python Wrapping, Python Scripted Modules
• Loadable Modules
• Editor
• Annotations (Evolution of Fiducials and
  Measurements)
• Many things

22
Slicer4 Whats Missing

• Many Core Modules
• Volume Rendering
• Hierarchies
• IGT Modules
• Extensions
• Progress / Status feedback
• Your Module!
• Please help identify what is needed to support
  your work.

23
Integration Options
24
Extensions

• Keep the base package lean and mean
• Modules have individual identity
• Per-module web site, svn, downloads, mailing
  lists, wiki
• Allow users to assemble their own set of tools
• Customized Bundles by task or application
• Easy to download compatible extensions
• Analogous to Firefox extensions
• Integrate extension builds into
  developer/nightly/release processs
• NITRC Supplement to NA-MIC helping to pay for
  needed infrastructure (Neuroimaging Informatics
  Tools and Resources Clearinghouse)
• NITRC can host neuroimaging projects (gforge
  implementation)

25
Extensions

• Keep the base package lean and mean
• Modules have individual identity
• Per-module web site, svn, downloads, mailing
  lists, wiki
• Allow users to assemble their own set of tools
• Customized Bundles by task or application
• Easy to download compatible extensions
• Analogous to Firefox extensions
• Integrate extension builds into
  developer/nightly/release processs
• NITRC Supplement to NA-MIC helping to pay for
  needed infrastructure (Neuroimaging Informatics
  Tools and Resources Clearinghouse)
• NITRC can host neuroimaging projects (gforge
  implementation)

26
Extensions Interface
27
Slicer4 Radiation Therapy Topics

• DICOM RT
• Greg Sharp will Describe Slicer3/Slicer4
  Plastimatch I/O
• ITKv4 was Planning DICOM RT Support
• DCMRT is a DCMTK Extension from Oldenberg (Offis)
  and DKFZ in Heidelberg
• Multi-Structure Volume Representation in VTK and
  ITK
• Vector Valued Pixel Data
• Efficient Transforms and Rendering
• Preserving DICOM Header Information
• CUDA/OpenCL Deployment (shadie)
• Server-Side-Slicer
• In Prototype Phase
• Web Apps for Deployment to Workstations via
  Browser
• Javascript, HTML5, WebGL