Engineering:Project:ITK Tract Analysis

From NAMIC Wiki
Jump to: navigation, search
Home < Engineering:Project:ITK Tract Analysis

Project: Clustering and Analysis of DTI Fibers.

File:Namic-tractanalysis.ppt

Goals:

  1. Develop ITK modules for clustering DTI fibers from tractography. Using Normalized Cut algorithm of Shi and Malik. Improve efficiency.
  2. Collaborate with Slicer Tractography group to get perspective on different approaches to tractography and fiber tract analysis.
  3. Make UNC tools compatible with other NA-MIC tools
    1. Read NA-MIC NRRD DTI data
    2. Develop a tool that can convert ITK DTITubeSpatialObject files to vtkPolyData files and vice-versa. The UNC tools currently use the ITK files, while other tools use vtk files.
  4. Develop a generic tensor statistics computation module for fiber tract analysis. Currently implemented for and integrated into FiberViewer. Collaborate with Tom Fletcher.

Current Status:

  1. Working ITK implementation of clustering algorithm that is generic over objects with defined distance measure.
  2. Tensor statistics implemented in FiberViewer, but not yet a generic module.
  3. Tensor field computation for NRRD DTI data via Xiadong's NrrdDTIReconstruction tool.

Current Tools

  1. Fiber Tracker - Tool for tracking fibers from DTI data. http://www.ia.unc.edu/dev/download/fibertracking/index.htm
  2. Fiber Viewer - Tool for fiber editing, clustering, analysis. (Normalized Cut based clustering goes here) http://www.ia.unc.edu/dev/download/fiberviewer/index.htm
  3. SNAP - Used for defining ROI used in tracking. (included in InsightApplications).
  4. Data conversion tool: from nrrd to gipl/analyze.
  5. Fiber format conversion tool: from UNC fiber to vtk polydata.

Test Data: Currently using in-house data. A sample DTI dataset is available here in GIPL format http://www.ia.unc.edu/dev/download/fibertracking/DTI_Test.zip. Also testing on Brockton data from BIRN repository.

Team Members:

  • Casey Goodlett - UNC
  • Xiaodong Tao - GE
  • Isabelle - UNC
  • Mahnaz - MIT
  • Andrea - BWH