2009 Winter Project Week:LesionAlgorithms

Return to Project Week Main Page

Key Investigators

Mark, Marcel, Vince, Ross, Jeremy

Objective

We have tested several algorithms and approaches for classifying white matter lesions.
There are also other methods within the NA-MIC kit and and external approach that is being developed to be used within Slicer (WMLS).
Currently for our end to end tutorial, we have implemented a novel algorithm because the generalizability of our previous method to the clinical sample was poor.
We need to make a final selection on choice of algorithm and any resulting optimization strategy to be used for final analysis and a timely manuscript.

Approach, Plan

Review comparisons of NA-MIC kit tools EM-Segment, Prastawa Method, Magnotta Method, a new method a UPenn, etc
What is the best approach to use for the roadmap project?
What if anything can be leveraged or combined into a universal method?
This work should lead towards planned MICCAI (march 8 due date) or other venue paper on white matter lesion also plan on immediately after full application/analysis of a clinical lupus sample (n=62).

Progress

• We have prepared an HBM abstract for submission on Sunday, 01/11/09
• Scoped out a manuscript submission to MICCAI for March 8th.
• In a parallel effort we'll be doing a comparison of lesion classification methods for neuroimage submission with Marcel, Vince Magnotta, and other intersted parties for future submission at a conference to be determined.