MIT CSAIL Biomedical Image Analysis Group

1
MIT CSAIL Biomedical Image Analysis Group

• Polina Golland, Sandy Wells, Eric Grimson

2
What we contribute

• Segmentation methods
• Especially shape and atlas based methods
• Statistical analysis of shapes
• Anatomical structures
• DTI tracts
• Registration
• Large scale population co-registration
• fMRI analysis
• Anatomically guided detection
• Population analysis

3
Segmentation

• Atlas- and shape- based segmentation
• Integrated into NAMIC-kit
• Collaborative studies with DBPs
• Publications
• Pohl et al. MICCAI 2005, MICCAI 2006, NeuroImage
  2006
• Koo et al. AGP 2006 clinical application
• In preparation clinical paper with Harvard

4
DTI

• Clustering of fiber tracks
• Atlas construction
• Segmentation
• Mature components are in NAMIC-kit
• Collaborative studies with DBPs
• Publications
• ODonnell et al. MICCAI 2005, MICCAI 2006, ISMRM
  2006
• Ziyan et al. MICCAI 2006, Maddah et al. MICCAI
  2005, ISBI 2006
• ODonnell et al. AJNR 2006 clinical application

5
Registration

• Information-theoretic group-wise registration
• Integration into NAMIC-kit in progress
• Collaborative studies in progress
• Publications
• Zollei et al. ICCV Workshop 2005 (Best paper
  award), WBIR 2006

6
fMRI Analysis

• Anatomically guided fMRI detection
• Functional hierarchies
• Integration into NAMIC-kit in progress
• Demonstrated on DBP data
• Publications
• Ou Golland, IPMI 2005
• In preparation TMI article, IPMI 2007, MICCAI
  2007

7
What we get from NAMIC
NAMIC provides a collaborative base for
interaction with driving applications

• Amplifies computational work we are doing under
  core NIH RO1
• Complements medical robotics work as part of NSF
  ERC with JHU
• Serves as a main avenue of collaborations in
  clinical neuroscience