Difference between revisions of "2011 Winter Project Week:NAMICShapeAnalysis"

From NAMIC Wiki
Jump to: navigation, search
Line 15: Line 15:
  
 
<h3>Objective</h3>
 
<h3>Objective</h3>
Statistical shape analysis methods have emerged within the last decade to allow for a localized analysis of shape.The UNC shapeAnalysisMANCOVA pipeline is a unified method for local shape analysis that can accomodate different number of variates and contrasts. Unlike current shape analysis frameworks,it also allows to include any number of associated variables in the statistical analysis of the data.
+
Statistical shape analysis methods have emerged within the last decade to allow for a localized analysis of shape.The UNC shapeAnalysisMANCOVA pipeline is a unified method for local shape analysis that can accommodate different number of variates and contrasts. Unlike current shape analysis frameworks,it also allows to include any number of associated variables in the statistical analysis of the data.
 
This tool has been designed to interact seamlessly with the existing UNC SPHARM-PDM based shape analysis toolbox. Indeed, the point-based models computed with the SPHARM-PDM tool can be used in combination with this pipeline to perform quantitative morphological assessment of structural changes at specific locations.
 
This tool has been designed to interact seamlessly with the existing UNC SPHARM-PDM based shape analysis toolbox. Indeed, the point-based models computed with the SPHARM-PDM tool can be used in combination with this pipeline to perform quantitative morphological assessment of structural changes at specific locations.
 
The aim is to  improve this UNC pipeline by integrating the Particle-based Shape Correspondence, and using SPHARM as an initialization.
 
The aim is to  improve this UNC pipeline by integrating the Particle-based Shape Correspondence, and using SPHARM as an initialization.
Line 41: Line 41:
 
doing the permutations tests (considering the linear varibles).
 
doing the permutations tests (considering the linear varibles).
  
The data scenes (MRML scene) are created and can be displayed within 3D Slicer. <foo>.
+
The data scenes (MRML scene) are created and can be displayed within 3D Slicer.
  
Our plan for the project week is to first try out to find a path between Boost and ITK, VTK, Slicer lib. <bar>,...
+
Our plan for the project week is to:
 +
* try to get rid of boost library - find similar method in Slicer libraries if possible
 +
* use GenerateCLP to integrate shapeAnalysisMANCOVA_Wizard directly within Slicer3
 +
* Load MRML scenes created for quality control directly within Slicer
 +
* improve the shapeAnalysisMANCOVA_Wizard
  
 
</div>
 
</div>

Revision as of 16:23, 17 December 2010

Home < 2011 Winter Project Week:NAMICShapeAnalysis

Key Investigators

  • UNC: Lucile Bompard, Clement Vachet, Martin Styner
  • Utah: Chris Gloschat

Objective

Statistical shape analysis methods have emerged within the last decade to allow for a localized analysis of shape.The UNC shapeAnalysisMANCOVA pipeline is a unified method for local shape analysis that can accommodate different number of variates and contrasts. Unlike current shape analysis frameworks,it also allows to include any number of associated variables in the statistical analysis of the data. This tool has been designed to interact seamlessly with the existing UNC SPHARM-PDM based shape analysis toolbox. Indeed, the point-based models computed with the SPHARM-PDM tool can be used in combination with this pipeline to perform quantitative morphological assessment of structural changes at specific locations. The aim is to improve this UNC pipeline by integrating the Particle-based Shape Correspondence, and using SPHARM as an initialization.




Approach, Plan

The developed shapeAnalysisMANCOVA is a C++ program that can be run both with shapeAnalysisMANCOVA_Wizard as well as a command line tool. In order to compile shapeAnalysisMANCOVA, ITK, VTK, GenerateCLP and Boost Libraries for C (v 0.39.1) must already be installed. The point-based models will be analyzed with methods using multivariate analysis of covariance (MANCOVA). The steps performed in order to obtain the statistical analysis are the following, for each iteration:

computing a General Linear Model (GLM) to test group differences at every surface location,

doing the Metric computation thanks to the Multivariate analysis of covariance (MANCOVA),

computing the P-values and controlling for the mutliple testing problem,

doing the permutations tests (considering the linear varibles).

The data scenes (MRML scene) are created and can be displayed within 3D Slicer.

Our plan for the project week is to:

  • try to get rid of boost library - find similar method in Slicer libraries if possible
  • use GenerateCLP to integrate shapeAnalysisMANCOVA_Wizard directly within Slicer3
  • Load MRML scenes created for quality control directly within Slicer
  • improve the shapeAnalysisMANCOVA_Wizard

Progress


Delivery Mechanism

This work will be delivered to the NA-MIC Kit as a

  1. ITK Module
  2. Slicer Module
    1. Built-in
    2. Extension -- commandline YES
    3. Extension -- loadable YES
  3. Other shapeAnalysisMANCOVA_Wizard