1) The document discusses what makes a successful open-source medical imaging project. It examines projects like Mango/Papaya, XMedCon, ImageJ, XNAT, and Kitware that have been in operation for many years. 2) Some attributes of successful projects identified are having a passionate leader, addressing an unmet need, producing high-quality work, utilizing existing resources, rewarding contributors, and continual improvements. 3) The document provides advice for starting a new open-source project, such as identifying a product or service, choosing a name, building an online presence, and soliciting feedback.

1. What Makes a Successful
Open-Source Project in
Medical Imaging?
OrthancCon I, 2019
Andrew Crabb
The Johns Hopkins University
I Do Imaging

2. I Do Imaging
The searchable database of free and
open-source medical imaging
software
In operation 20 years
330 active medical imaging projects
12,000 monthly visitors
30,000 subscribers

3. Successful Open-Source Medical Imaging
Project
Successful
Open Source
Medical Imaging
Project
Project
Typically but not always a software project
Other forms: Website, blog, database
Open Source
Used loosely to mean free applications or services
Some good projects don’t release source code
Sometimes open source is required (license, funding)
Medical Imaging
Constraint here: Only projects that involve images
Not addressing informatics, HIS, HL7, FHIR
Successful? You decide – it’s your project
What is your goal? Solve a problem, new skills/field,
recognition, entrepreneurship, share knowledge
What is it? Software, service, website, blog, channel
What does it do? Image display, analysis, informatics,
networking, utility, library
Problem domain? Clinical/research, technique,
modality, image format, platform or language
Success could just be the clarification of your own
ideas, to be implemented in your next project

4. Five Successful Projects
Problem Size
Solution
Size
Mango/Papaya

5. XMedCon (Erik Nolf)
Medical image conversion, emphasis on
nuclear medicine file formats
18+ years of updates
Oriented to research and command line
Includes other code (NIfTI, DICOM, GIF)
Open source and binary distributions
Available for many platforms

6. XMedCon

7. Mango/Papaya (Univ of Texas)
Product of one neuro imaging lab
Funded project (US NIH)
Mango: Java neuroimaging app
Research image viewing and analysis
Wide range of file formats
Command line integration and scripting
Plugin architecture with extensions from the
neuroimaging community

9. ImageJ (US NIH)
Long-lasting Java image analysis app
Targeted audience (image processing,
optical imaging)
500 plugins available
Automate with 300 available macros
Loyal, active user base

10. XNAT (Washington University St Louis)
Research/clinical neuroimaging
informatics platform since 2004.
REST interface
Many related tools for uploading and
editing DICOM files
De-identification language
Extensible, with plugin marketplace

11. Kitware
Commercial company
Started with one product (VTK)
Influential publisher
Basis of many projects

12. Attributes of Successful Projects
Passion for the topic
• Most projects stop through lack of
interest or time
Identifying a market
• Address an underserved need
• Niche/smaller use cases
• New technologies
A quality product
• Competing for time and attention
• Even free software must be good
Use existing resources
• Software libraries
• Frameworks (web apps and text)
Reward helps
• Funding for public benefit
• Reputation and recognition
Continual progress
• Abandoned projects will be ignored or
supplanted
Community contact
• Publications, social media, workshops,
conferences
• Release notes, issue tracker
A good user experience
• Memorable name
• Attractive and responsive website
• Low barrier to adoption

13. Starting Your Own Project
You have something to contribute
Identify your product
Think of a name
Build your brand – domain name, social media
Improve
your
product
PublishMetrics /
feedback

14. Free Resources

15. Free Resources – Amazon Web Services

16. Free Resources – Microsoft Azure

17. Orthanc
Passion for the topic
• Most projects stop through lack of
interest or time
Identifying a market
• Address an underserved need
• Niche/smaller use cases
• New technologies
A quality product
• Competing for time and attention
• Even free software must be good
Use existing resources
• Software libraries
• Frameworks (web apps and text)
Reward helps
• Funding for public benefit
• Reputation and recognition
Continual progress
• Abandoned projects will be ignored or
supplanted
Community contact
• Publications, social media, workshops,
conferences
• Release notes, issue tracker
A good user experience
• Memorable name
• Attractive and responsive website
• Low barrier to adoption