EMSegmentation in Slicer 3

1
EMSegmentation in Slicer 3
B. Davis, S. Barre, Y. Yuan, W. Schroeder, P.
Golland, K. Pohl
2
Overview

• Introduction
• EM Module Step-By-Step
• Feedback Discussion
• Live Demo

3
Motivation
4
Hierarchical Tree
5
Applications of EM Segmenter
Subcortical SegmentationPsychiatry Neuroimaging
LaboratoryBWH, Harvard
White Matter LesionCenter for Neurological
Imaging BWH, Harvard
6
Goals

• Design automatic segmenter that
• is easy to use
• adapts to variety of scenarios
• works on large data sets
• is a research tool

Slicer2
Slicer3
7
Overview

• Introduction
• EM Module Step-By-Step
• Feedback Discussion
• Live Demo

8
Wizard Interface
Parameter Set
Tree
Atlas
Target
Intensity

• Separates complex tasks into a sequence of
  simpler steps
• Checks user input before each transition
• Provides consistent access to help

Parameters
Registration
Run
9

• Parameter set defines segmentation scenario
• Atlas, Images, Algorithm parameters

Parameter Set
Tree
Atlas

• Create new parameter set
• Apply/modify existing parameter set

Target
Intensity
Parameters
Registration
Run
10
Defines a hierarchy of anatomical structures
Parameter Set
Tree
Atlas
Target
Intensity
Parameters
Registration
Run
11
Assign atlas to anatomical structures
Parameter Set
Tree
Atlas
white matter
csf
Target
Intensity
grey matter
background
Parameters
Registration
Run
12
Choose input channels
Parameter Set
Tree
Atlas
Target
T1
Intensity
Parameters
Registration
Run
T2
13
Define intensity distribution for each structure
Parameter Set
Tree
Atlas
Target
Intensity
Parameters
Registration
Run
14
Specify node-based segmentation parameters
Parameter Set

• Influence of
• Input channels
• Atlas
• Smoothing
• Relative weight to other structures
• Stopping conditions

Tree
Atlas
Target
Intensity
Parameters
Registration
Run
15
Specify atlas-to-input channel registration
Parameter Set
Tree
Atlas
white matter
csf
Target
Intensity
grey matter
background
Parameters
Registration
Run
T2
T1
16
Segment input channels using parameters
Parameter Set
Tree
Atlas
Target
Intensity
Parameters
Registration
Run
17
Pipeline
2
3
1
AtlasAlignment
EMSegmentation
IntensityNormalization
18
EM Segmenter
Image
Prior
Hierarchy
Labelmap
19
Level 1
Prior Information
20
Level 2
IMAGE
ICC
Current Parameter
ROI
21
Example Tree
22
(No Transcript)
23
Overview

• Introduction
• EM Module Step-By-Step
• Feedback Discussion
• Live Demo

24
Resouces

• Slicer3 EMSegment Wiki page http//wiki.na-mic.o
  rg/Wiki/index.php/Slicer3EM
• Project Description
• Steps in EMSegment Workflow
• Future Work
• Implementation Details
• EMSegment Tutorial
• Slicer2 Material
• Tutorial http//wiki.na-mic.org/Wiki/index.php/S
  licerWorkshopsUser_Training_101
• Publications
• K.M. Pohl , S. Bouix, R. Kikinis, W.E.L. Grimson,
  Anatomical Guided Segmentation with
  Non-Stationary Tissue Class Distributions in an
  Expectation-Maximization Framework, In Proc. ISBI
  2004, pp. 81-84,2004
• K.M. Pohl, S. Bouix, M.E. Shenton, W.E.L.
  Grimson, R. Kikinis, Automatic Segmentation Using
  Non-Rigid Registratio, short communications of
  MICCAI 2005

25
Feedback Discussion

• Priorities for future development
• Class overview panel
• Graphical Display
• Controlled vocabulary
• Library of Templates
• One-Step-Segmentation

26
Acknowledgements

• Steve Pieper
• Alex Yarmarkovich
• Wendy Plesniak
• Slicer developer community
• Psychiatry Neuroimaging Laboratory
• NAMIC

27
Acknowledgements

• Kitware Developer

28
Overview

• Introduction
• EM Module Step-By-Step
• Feedback Discussion
• Live Demo

29
Editing the Tree