MICCAI 2013 Tutorial - Common architecture for algorithm development and deployment
Introduction
Medical image computing research routinely requires integration with existing research workflows or clinical systems. Yet it can take years of development effort to build and debug the necessary support software. The Common Toolkit (CTK) initiative provides a shared code base to address these interoperability issues. Development of CTK is driven by representatives from several mature open source platforms who recognize the benefit of avoiding duplication and wish to share the results of their experience.
Scope
This tutorial will provide a comprehensive overview and in-depth presentations about the state-of-the-art in medical image computing software. Starting with an introduction to CTK, some of the major medical imaging software platforms using CTK will be presented, along with examples of clinical research supported by CTK code. Participants gain insight into the particular strengths and foci of each platform and how they can be leveraged to boost algorithmic research.
The main theme of the tutorial is support of algorithm development, deployment and interoperability. Attendees will learn how a few lines of code can make their medical image software compatible with a wide range of deployment environments. A substantial part will be the presentation of real world use cases and solutions, showing participants how to actually leverage the presented technologies.
Examples of what an attendee will learn include:
- How a single version of ITK registration code can be made compatible with half a dozen different research platforms
- How a CAD algorithm can be easily integrated with a clinical PACS system
- How to bundle MATLAB code for integration into a clinical workflow
Objectives
The tutorial has the following objectives:
- Present the state-of-the-art in medical image computing software platforms.
- Educate participants how to efficiently develop new algorithms, benefiting from the infrastructure provided by the different software platforms.
- Show how to easily re-use algorithms of other researchers.
- Foster translation of imaging algorithms to a clinical environment by demonstrating interoperability and deployment techniques.
Program
08:30 - 09:15 The Common Toolkit: What it provides and how to use it.
Sascha Zelzer, Steve Pieper09:15 - 10:00 Medical imaging platforms perspective: How platforms like 3D Slicer, MITK, NifTK and MedInria use CTK.
(Welcome, Intro, How to use CTK, Example code)
Steve Pieper, Marco Nolden, Matt Clarkson, Olivier Commowick10:30 - 11:30 Algorithm developer perspective: How to and why make your algorithm compatible with CTK. Interoperability use cases.
(Slicer, MITK, NifTK, medInria)
Sonia Pujol, Matt Clarkson, Hans Jonson, Sascha Zelzer, Steve Pieper11:30 - 12:30 Deployment scenarios: Classrooms, Research Labs, Reading Rooms, ORs, Workstations, Web Servers, Clusters and Clouds.
(Writing CLIs, NiftyReg use case, Matlab example, CLI wrap-up)
Matt Clarkson, Steve Pieper, Sascha Zelzer
(Workstation and Research Use, Reading Room and OR, DICOM Application Hosting)
Registration
Registrations are handled by the MICCAI registration system.
Organizers
Sascha Zelzer
German Cancer Research Center, Heidelberg
Email: s.zelzer@dkfz.de
Marco Nolden
German Cancer Research Center, Heidelberg
Email: m.nolden@dkfz.de
Sonia Pujol
Brigham and Women’s Hospital, Boston
spujol@bwh.harvard.edu
Steve Pieper
Isomics, Inc., Cambridge
pieper@isomics.com
Matt Clarkson
University College London, London
m.clarkson@ucl.ac.uk