Minutes 072007: Difference between revisions

From KitwarePublic
Jump to navigationJump to search
 
(One intermediate revision by the same user not shown)
Line 25: Line 25:
* Image Registration Optimization
* Image Registration Optimization
** Jim suggested considering an API in the Transform where the working memory for the Jacobian is passes to the Transform.
** Jim suggested considering an API in the Transform where the working memory for the Jacobian is passes to the Transform.
** Hans suggested to have the Transform create its own working memory, so that the Metrics don't need to know about the internal structure of the Transform.
** Jim suggested to create an API where the Transform use the Jacobian directly to compute the matrix product with the Image gradient.


== Action Items ==
== Action Items ==

Latest revision as of 18:10, 20 July 2007

Attendees

  • Jim Miller
  • Ken Williams
  • Hans Johnson
  • Nick Tuiston
  • Alex Gouaillard
  • Luis Ibanez

Discussion

  • Approved CVS write access to Rogue Research developers
  • ImageIOSteaming
    • Kent, Hans: gave an update on the Nifti library changes related to ImageIO Streaming.
    • How to define the API of a streamable writer?.
    • How to manage compression?.
    • Dialog between ImageFileWriter and ImageIO class for passing a SplitterImageRegion.
    • Put back the CanStream() method in the ImageIO, and just not call it from the ImageFileReader to keep backward compatibility
    • Rename CalculateStreamableReadRegionFromRequestedRegion() as GenerateStreamableReadRegionFromRequestedRegion()
  • itkQuadMesh Status
    • Alex gave an update on the status of the itkQuadMesh
    • We agreed on moving a second batch of classes into the Review directory.
    • Alex will use the itkVTKPolyDataReader class in the Code/Review directory for reading VTK surfaces to be used as input in the tests.
    • Jim pointed out some naming conventions to follow.
  • Image Registration Optimization
    • Jim suggested considering an API in the Transform where the working memory for the Jacobian is passes to the Transform.
    • Hans suggested to have the Transform create its own working memory, so that the Metrics don't need to know about the internal structure of the Transform.
    • Jim suggested to create an API where the Transform use the Jacobian directly to compute the matrix product with the Image gradient.

Action Items

  • Move the ImageIOStreaming into ITK.
    • Adapt ImageSerieReader.