Skip to content

Latest commit

 

History

History
65 lines (43 loc) · 3.5 KB

File metadata and controls

65 lines (43 loc) · 3.5 KB

Back to Projects List

Segment Editor clinical workflows

Key Investigators

  • Csaba Pinter (Queen's University, Canada)
  • Till Best
  • Nikos Makris (Brigham and Women's Hospital and Harvard Medical School, USA)
  • Filippo Cicali (Massachusetts General Hospital and Harvard Medical School, USA)
  • Tina Kapur (Brigham and Women’s Hospital and Harvard Medical School, USA)

Participating

  • Marco Nolden (DKFZ, Germany)
  • Attila Nagy (SZTE Medical School, Hungary)
  • Sharon Peled (Brigham and Women's Hospital and Harvard Medical School, USA)

Project Description

Objective

  • Segment Editor is the new manual and semi-automated image segmentation module in Slicer. It has powerful tools, but the learning curve can be steep for segmenting more complex cases.
  • The objective is to help researchers aiming to use Segment Edtitor use it for their specific use cases, and augment the training material based on the results

Approach and Plan

  • Each participating researcher brings a dataset, for which we'll attempt to develop an effective segmentation workflow
  • Add the developed workflows in the Slicer Segmentation Recipes page
    • Publicly share dataset
    • Create a tutorial page for the segmentation case including screenshots

Progress and Next Steps

  • Insula segmentation in MRI with Filippo Cicali and Nikos Makris
    • Possible solution: Use Grow from seeds iteratively, then use scissors and masked paint to cut
    • Consider speeding up segmentation by using a stylus instead of a mouse, and explore faster ways of slice navigation (3D mouse, webcam tracking, etc?)
  • Brachytherapy needle segmentation in MRI for training deep learning models
    • Possible solution: Add "tube" mode from Markups to Models extension to Segment Editor, as an option to surface cut or a new effect in a new extension
    • Deep learning training support: need cohort segmentation slicelet that allows configuring segment editor before segmenting first patient for simpler, less distracting user interface for easier usage and smoother and more reliable segmentation workflow
  • Minor help to multiple projects involving segmentation

Illustrations

Whole heart segmentation:

Whole heart segmentation

Spine phantom 3D printing:

Spine phantom 3D printing

Segment Editor in presentation in Casa África:

Segment Editor in presentation in Casa África

Background and References