NA-MIC Project Weeks

Back to Projects List

Skin Surface Segmentation for NousNav

Key Investigators

• Reuben Dorent (Brigham and Women's Hospital, Harvard Medical School, USA)
• Colin Galvin (Brigham and Women's Hospital, Harvard Medical School, USA)
• Sarah Frisken (Brigham and Women's Hospital, Harvard Medical School, USA)
• Tina Kapur (Brigham and Women's Hospital, Harvard Medical School, USA)
• Sam Horvath (Kitware, USA)

Presenter location: In-person

Project Description

This project was initiated during the PW36. It aims to create an automated skin segmentation tool for pre-operative scans for NousNav.

A model has been already trained to automatically segment scans in multi-parametric MRI. In this project, we aim to integrate the developed tool into Slicer for further integration in NousNav.

Objective

1. Create a Slicer Extension for automatic skin segmentation based on a pre-trained nnUnet framework
2. Test the Slicer Extension in different settings, including different OS and different hardware configuration (only CPU, GPU, Mac)
3. Integrate the Slicer module into NousNav

Approach and Plan

1. Leverage the existing TotalSegmentator Slicer extension as template
2. Create a small database that can be shared to test the algorithm in different settings
3. Discuss with the NousNav development team for its integration

Progress and Next Steps

1. The Slicer extension has been implemented and is publicly available here.
2. The extension has been tested on CPU (MacOS and Windows) and GPU (Windows)
3. The integration in NousNav is still pending.

Illustrations

Background and References

No response