ivadomed / ivadomed/model-spinal-rootlets
Integrating brachial plexus & DRG segmentation
- Dominant language
- Python
- Stars
- 8
- Forks
- 2
- PR merge metrics
- No merged PRs in 30d
Description
Propose integrating a new model for automated bilateral segmentation of the brachial plexus and dorsal root ganglia at cervical levels C5–C8 from 3D STIR sequences
Dataset: approx. 400 total participants (340 on Siemens, 60 on GE), including 3 total sites
Model details:
- Architecture: Cascaded nnU-Net (2-stage, 5-fold ensemble)
- Stage 1: 3-class coarse segmentation (background / DRG / plexus)
- Stage 2: Fine 16-class bilateral segmentation (BP C5–C8 + DRG C5–C8, bilaterally [separate left & right])
Contributor guide
No contributing guide indexed for this repository
Research direction
Start by surveying the repository's existing model integration entry points and conventions, since the issue names no files or tests. Compare the proposed cascaded nnU-Net stages, label scheme, imaging data, and multi-site setup with current project support. Done means the new bilateral C5–C8 plexus and DRG model is integrated and its behavior is validated according to the repository's existing checks.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- computer-vision, machine-learning
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100