2017 Winter Project Week/Diffusely abnormal white matter segmentation with 3d U-net

From NAMIC Wiki
Jump to: navigation, search
Home < 2017 Winter Project Week < Diffusely abnormal white matter segmentation with 3d U-net

Key Investigators

  • Mohsen Ghafoorian (BWH, Radboud University)
  • Bram Platel (Radboud University)
  • Sandy Wells (BWH)
  • Tina Kapur (BWH)
  • Charles Guttmann (BWH)
  • Hans Meine (Univ. Bremen, Fh MEVIS)

Project Description

Objective Approach and Plan Progress and Next Steps
  • Diffusely abnormal white matter (“DAWM”) are fuzzy-bordered areas of subtly increased signal

intensity on proton density or T2-weighted images. These abnormalities have been referred to as dirty white matter or dirty-appearing white matter. The goal is to develop an automated system that accurately segments diffusely abnormal white matter.

  • We would like to use the 3D unet that is shown to be a great architecture for biomedical image segmentation.
  • The 3D-Unet architecture is implemented in Lasagne and it is currently being trained.
Retrieved from "https://www.na-mic.org/w/index.php?title=2017_Winter_Project_Week/Diffusely_abnormal_white_matter_segmentation_with_3d_U-net&oldid=96113"