NA-MIC Project Weeks

Back to Projects List

SlicerAIGT

Key Investigators

• Gabor Fichtinger (PerkLab, Queen’s)
• Andras Lasso (PerkLab, Queen’s)
• Kyle Sunderland (PerkLab, Queen’s)

Project Description

Artificial intelligence (AI) has promising applications in image-guided therapy (IGT), but significant infrastructure development is needed to implement such systems.

Objective

Develop tools that make it easy to train AI and use AI models in real-time applications.

Approach and Plan

Get information from project week attendees about their current AI workflows, shortcomings, and wishlist.

Progress and Next Steps

• SlicerAIGT repository contains collection of scripts, Jupyter notebooks, and Slicer utility modules: to be cleaned up and generalized
• Discussed medical image computing AI frameworks that may be considered to be supported:
  • NVidia Clara: already integrated with Slicer’s Segment Editor, NVidia-developed models are available on Nvidia Cloud
  • modelhub.ai: evaluated during this week
  • niftynet: model zoo contains pre-trained models