Difference between revisions of "2008 Winter Project Week:Astronomical Medicine"

From NAMIC Wiki
Jump to: navigation, search
 
(8 intermediate revisions by 2 users not shown)
Line 23: Line 23:
 
<h1>Approach, Plan</h1>
 
<h1>Approach, Plan</h1>
  
Our approach is summarized on the [http://astromed.iic.harvard.edu Astronomical Medicine web site].  The main challenge to this approach, at the moment, is the impedance mismatch between celestial coordinates and medical imaging coordinates, and a lack of infrastructure within ITK and Slicer to allow for alternative coordinate systems.
+
Our approach is summarized on the [http://astromed.iic.harvard.edu Astronomical Medicine web site].  The main challenge to this approach at the moment, is the impedance mismatch between celestial coordinates and medical imaging coordinates, and a lack of infrastructure within ITK and Slicer to allow for alternative coordinate systems.
  
Our plan for the project week is to implement a mechanism that will allow world coordinate system information to travel through the ITK pipeline via an ITK image's metadata dictionary.  If we accomplish this, and there is time left over, then we hope to implement bidirectional procedural transforms to map between voxel coordinates and world coordinates.
+
Our plan for the project week is to make further progress in integrating our ITK FITS reader into Slicer as a plug-in, and to begin work on Slicer GUI enhancements to support our FITS plug-in and to allow the display of astronomical coordinates in slice viewers.
 
</div>
 
</div>
  
Line 31: Line 31:
  
 
<h1>Progress</h1>
 
<h1>Progress</h1>
 +
====January 2008 Project Week====
 +
 +
* Created vtkCoordinateSpaceManager, a container that manages multiple named coordinate spaces and the transforms between them. A reader can populate a coordinate space manager with all the coordinate spaces it knows about.  Registration tools can link otherwise unrelated coordinate spaces, and the manager can hand back a (possibly composite) transform between any two related spaces. This module hasn't been incorporated into Slicer3 yet.
 +
* With Wendy Plesniak's help, we now have the ability to recenter on data not located at the origin in Slicer. She's incorporating similar changes into the "look from" Slicer control. We'll need to look at autozooming as well for very small volumes.
 +
* Slicer3 volume rendering works reasonably well with astronomical data, but the world size of model seems to influence rendering appearance. Solved some performance issues on the Mac.
 +
* Experimental linearized celestial coordinate system implementation being tested now.
 +
* Prototyped Slicer3 plugin module adapter for fits2itk.
 +
** Debugged performance issues with the Slicer3 implementation of plugin modules for OS X and submitted bug report.
 +
* Prototyped ITK-based NaN masking filter.
 +
** Still need standard way to deal with NaNs in ITK/VTK/Slicer3.
 +
 
====June 2007 Project Week====
 
====June 2007 Project Week====
  
Worked with Jim Miller to implement preservation of the ITK metadata
+
Worked with Jim Miller to implement preservation of the ITK metadata dictionary through the Slicer pipeline, worked with Luis Ibanez to help implement ITK streaming file I/O, set up more robust dependency rules in our CMake config, and debugged a problem with our use of ITK filters.
dictionary through the Slicer pipeline, worked with Luis Ibanez to
 
help implement ITK streaming file I/O, set up more robust dependency
 
rules in our CMake config, and debugged a problem with our use of ITK
 
filters.
 
  
 
====Jan 2007 Project Half Week====
 
====Jan 2007 Project Half Week====
  
Solved problem of itk::ImageIOBase being incompatible with CFITSIO's
+
Solved problem of itk::ImageIOBase being incompatible with CFITSIO's "extended filename syntax", identified compatibility issues present in the most recent version of Slicer 3, and devised a plan for passing FITS WCS transforms around in Slicer by encapsulating them in itk::Transform objects held in an image's metadata dictionary.
"extended filename syntax", identified compatibility issues present in
 
the most recent version of Slicer 3, and devised a plan for passing
 
FITS WCS transforms around in Slicer by encapsulating them in
 
itk::Transform objects held in an image's metadata dictionary.
 
  
 
====June 2006 Project Week====
 
====June 2006 Project Week====
  
Implemented a prototype FITS reader for ITK.  We later released this
+
Implemented a prototype FITS reader for ITK.  We later released a subsequent version as an open-source FITS-to-NRRD file converter, which is available [http://astromed.iic.harvard.edu/FITS-reader here].
as an open-source FITS-to-NRRD file converter, which is available
 
[http://astromed.iic.harvard.edu/FITS-reader here].
 
  
 
</div><br style="clear: both;" /></div>
 
</div><br style="clear: both;" /></div>
Line 57: Line 58:
 
<h1></h1><br>
 
<h1></h1><br>
 
====References====
 
====References====
* [http://am.iic.harvard.edu Astronomical Medicine web site]
+
* The [http://am.iic.harvard.edu Astronomical Medicine web site]
 
* [http://am.iic.harvard.edu/FITS-reader Open-source release] of our ITK-based FITS file reader/converter.
 
* [http://am.iic.harvard.edu/FITS-reader Open-source release] of our ITK-based FITS file reader/converter.
  

Latest revision as of 21:34, 5 June 2008

Home < 2008 Winter Project Week:Astronomical Medicine
Mockup of a future version of Slicer with features desired by the Astronomical Medicine project.


Key Investigators

  • Havard IIC: Douglas Alan
  • BWH/Harvard IIC: Mike Halle
  • Isomics: Steve Pieper

Objective

We are developing tools and enhancements to ITK and Slicer in order to adapt Slicer to be a tool that is useful for the visualization of astronomical data, specifically spectral line data cubes for now.

Approach, Plan

Our approach is summarized on the Astronomical Medicine web site. The main challenge to this approach at the moment, is the impedance mismatch between celestial coordinates and medical imaging coordinates, and a lack of infrastructure within ITK and Slicer to allow for alternative coordinate systems.

Our plan for the project week is to make further progress in integrating our ITK FITS reader into Slicer as a plug-in, and to begin work on Slicer GUI enhancements to support our FITS plug-in and to allow the display of astronomical coordinates in slice viewers.

Progress

January 2008 Project Week

  • Created vtkCoordinateSpaceManager, a container that manages multiple named coordinate spaces and the transforms between them. A reader can populate a coordinate space manager with all the coordinate spaces it knows about. Registration tools can link otherwise unrelated coordinate spaces, and the manager can hand back a (possibly composite) transform between any two related spaces. This module hasn't been incorporated into Slicer3 yet.
  • With Wendy Plesniak's help, we now have the ability to recenter on data not located at the origin in Slicer. She's incorporating similar changes into the "look from" Slicer control. We'll need to look at autozooming as well for very small volumes.
  • Slicer3 volume rendering works reasonably well with astronomical data, but the world size of model seems to influence rendering appearance. Solved some performance issues on the Mac.
  • Experimental linearized celestial coordinate system implementation being tested now.
  • Prototyped Slicer3 plugin module adapter for fits2itk.
    • Debugged performance issues with the Slicer3 implementation of plugin modules for OS X and submitted bug report.
  • Prototyped ITK-based NaN masking filter.
    • Still need standard way to deal with NaNs in ITK/VTK/Slicer3.

June 2007 Project Week

Worked with Jim Miller to implement preservation of the ITK metadata dictionary through the Slicer pipeline, worked with Luis Ibanez to help implement ITK streaming file I/O, set up more robust dependency rules in our CMake config, and debugged a problem with our use of ITK filters.

Jan 2007 Project Half Week

Solved problem of itk::ImageIOBase being incompatible with CFITSIO's "extended filename syntax", identified compatibility issues present in the most recent version of Slicer 3, and devised a plan for passing FITS WCS transforms around in Slicer by encapsulating them in itk::Transform objects held in an image's metadata dictionary.

June 2006 Project Week

Implemented a prototype FITS reader for ITK. We later released a subsequent version as an open-source FITS-to-NRRD file converter, which is available here.



References