EM Segmenter

Key Investigators

• Sylvain Jaume, MIT
• Nicolas Rannou, BWH
• Steve Pieper, BWH
• Ron Kikinis, BWH
• Polina Golland, MIT
• Koen Van Leemput, MGH

Objective

We re-visit the EM Segmenter code and algorithm to provide the clinician with an intuitive and robust tool to segment brain MRIs.

Approach, Plan

We are improving the EM Segmenter algorithm in three steps. First we are identifying the pieces of the EM Segmenter that are unintuitive to the clinician. Second we are designing a user interface that allow the clinician to use his/her knowledge to obtain the most reliable segmentation. Finally we are identifying the bottlenecks and re-writing the code to make the segmentation speed acceptable in a clinical setting.

Progress

We created a Slicer3 module for MRI Bias Field Correction as a pre-processing step for EM segmentation.