Edit this page

NA-MIC Project Weeks

Back to Projects List

CNN based Brain Masking Module

Key Investigators

Project Description

Develop a deep learning based Brain Masking Module with improved performance and accuracy over current alternatives.


  1. Objective A. Test accuracy and reliability of a Deep Learning based Brain Masking Solution
  2. Objective B. Integrate the solution into 3D slicer
  3. Objective C. Test and improve performance of the integrated solution

Approach and Plan

  1. Explore Image Registration options (EasyReg/MERMAID)
  2. Research current Deep Learning based Brain Masking (HD-BET, Auto Net)
  3. Get access to data for training and testing
  4. Create the solution
  5. Figure out how to integrate the solution
  6. Evaluate the performance

Progress and Next Steps

  1. Applied for NIH Dataset Request
  2. Tested HD-BET Segmentation
  3. Extracted PyTorch parameter(s) from HD-BET
  4. Begin building Slicer Module with HD-BET parameters


Swiss Skull Stripper SSS3D SSS4Spread

HD-BET Fast segmentation HDBETFast3D HDBETFast4Spread

HD-BET Accurate segmentation (5 model ensemble) HDBETAcc3D HDBETAcc4Spread

Background and References

EasyReg | HD-BET | Auto Net

Anatomical Data Augmentation via Fluid-based Image Registration Zhengyang Shen, Zhenlin Xu, Sahin Olut, Marc Niethammer. MICCAI 2020.

Region-specific Diffeomorphic Metric Mapping Zhengyang Shen, François-Xavier Vialard, Marc Niethammer. NeurIPS 2019.

Networks for Joint Affine and Non-parametric Image Registration Zhengyang Shen, Xu Han, Zhenlin Xu, Marc Niethammer. CVPR 2019.

Isensee F, Schell M, Tursunova I, Brugnara G, Bonekamp D, Neuberger U, Wick A, Schlemmer HP, Heiland S, Wick W, Bendszus M, Maier-Hein KH, Kickingereder P. Automated brain extraction of multi-sequence MRI using artificial neural networks. Hum Brain Mapp. 2019; 1–13. https://doi.org/10.1002/hbm.24750

S. S. Mohseni Salehi, D. Erdogmus and A. Gholipour, “Auto-Context Convolutional Neural Network (Auto-Net) for Brain Extraction in Magnetic Resonance Imaging,” in IEEE Transactions on Medical Imaging, vol. 36, no. 11, pp. 2319-2330, Nov. 2017, doi: 10.1109/TMI.2017.2721362.