Page History

• MIDRC - The challenge has ended and there were 4 finalists
• Test phase: 4 participants
• Validation phase: 7 participants
• We have algorithm descriptions
• CodaLab (Open Source Team):
  • We got the GPU worker working. I had deployed one temporarily and the Codalab team also created one and got it to work!
  • Automated tests now pass! (TACTIC: https://github.com/codalab/codabench)Image Removed
  • Check in any code changes as needed to fix issues found. We are moving into a phase where we are getting users to use Codabench. 
    • FAIR UNIVERSE – A group of people organizing competitions. This is good practice for testing Codabench.
  • codabench-test.lri.fr/competitions/91/
  • DGM: https://codabench-test.lri.fr/competitions/92/
• GPU submissions work
• I have 2 challenges launched preliminarily on codabench:

• • Federated learning challenge:
    • We still need to run the final two participant submissions but that will have to wait till November.
  • Image Comparator App:
    • Finished adding authentication to the app. Users can sign up and be signed in, log in and log out
• Met with below groups on future directions of their challenges (we meet again in two-three weeks to check progress):
  • DGM-Image Statistics challenge: Deep Generative Modeling for Learning Medical Image Statistic
  • TACTIC: Understanding Time-Activity Curve and Time-Integrated Activity Variations in Radiopharmaceutical Therapy Challenge
MIDRC:Challenge is going well. 4-5 active users making sample submissions out of 35.

• Federated Learning Challenge: 
  • Finalize runs on MGH data and check in with team for any retrospective. 
• CodaLab Codalab\Codabench (Open Source Team): 
  • Reviewing a teammate's code.
  • Next step is to fix a competition creation bug where organizers can make challenges using the UI.
• MIDRC:
  • Validation phase is beginning and we will see what that brings
• • Working on a bug with the manual competition creator
• Challenges Emily brought to the table (Meetings beginning in early Oct and they run roughly from Jan\Feb - May):
  • I still need to review a bit more but essentially they will have their own custom leaderboard and we will pre-build docker images for them to make sure they build fine
  • Then we will scp the images to their computers for analysis.
  • I finished understanding challenge details and am making the scoring program
  • DGM-Image Statistics challenge: Deep Generative Modeling for Learning Medical Image Statistic
  • TACTIC: Understanding Time-Activity Curve and Time-Integrated Activity Variations in Radiopharmaceutical Therapy Challenge
• MIDRC:
  • Not sure if we have more plans yet but I think there will be new challenge ideas
  Image Annotion App updates:
  • Adding the concept of users
  • Creating a cleaned up image upload, app image list creation and task creation functionality set tied to the user.
  • More updates to come including:
    
  • easy to use tools creation interface or functionality
  • possible integration with OHIF
  • MonaiLabel integration for better segmentation control.
  updated bootstrap (style)

01 Sep 13 Oct 2022 

15 Sep 2022 

29 Sep 27 Oct 2022