Difference between revisions of "Slicer3:3.6 Final Issues"

From Slicer Wiki
Jump to: navigation, search
Line 21: Line 21:
 
***After that (persistent-place, pick-once stealing mode, fid moved, and switch to default transform), all mouse modes work as expected.  
 
***After that (persistent-place, pick-once stealing mode, fid moved, and switch to default transform), all mouse modes work as expected.  
 
***The problem begins again if you delete all fiducials AND fiducial lists.
 
***The problem begins again if you delete all fiducials AND fiducial lists.
 
+
**Discussed with nicole -- having a look at the code.
 
*The diffusion tensor estimation module is using defaults that are not good and partially incorrect. Can the defaults be as follows:
 
*The diffusion tensor estimation module is using defaults that are not good and partially incorrect. Can the defaults be as follows:
 
**for Output DTI volume: DTI Volume
 
**for Output DTI volume: DTI Volume

Revision as of 18:05, 9 June 2010

Home < Slicer3:3.6 Final Issues

Final punch list for slicer 3.6 release:


  • VTK GPU ray-cast clipping issue (Julien says a fix is close)
  • new default color table (Nicole and Mike)
    • Mike working on creating a LUT from this hierarchy this subset, by Jun 7
    • Steve and Nicole to meet June 9 to integrate default editor color table
  • compare view + crosshair issue when loading a scene, the crosshair does not appear in all viewers until the slice location is changed, (Jim)
  • fiducial behavior/3D viewer behavior: (Steve, Nicole, Wendy to meet Jun 7)
    • tractography updates slow when drag stops (Steve)
    • saw one example of persistent place mode being stolen by pick mode. I think this has to do with inappropriate initialization of modes. It is possible for pick mode to take over until transient-pick mode is explicitly chosen.
    • To duplicate:
      • start with no fiducials or lists:
      • choose persistent-place.
      • place a fid, move mouse ever-so-slightly, hand appears.
      • If you do NOT press left-mouse-down, but move away, hand goes away and we remain in persistent-place mode.
      • place another fiducial and repeat by moving mouse every-so-slightly, hand appears again.
      • if you DO press left-mouse-down and then left-mouse-up WITHOUT moving mouse while hand is displayed, slicer switches to pick mode and STAYS in pick mode.
      • if you DO press left-mouse-down and then move mouse and drag fid before mouse-up, we switch into transform mode as expected.
      • After that (persistent-place, pick-once stealing mode, fid moved, and switch to default transform), all mouse modes work as expected.
      • The problem begins again if you delete all fiducials AND fiducial lists.
    • Discussed with nicole -- having a look at the code.
  • The diffusion tensor estimation module is using defaults that are not good and partially incorrect. Can the defaults be as follows:
    • for Output DTI volume: DTI Volume
    • for Output Baseline Volume: Baseline DTI Volume
    • for Otsu Threshold Mask: Otsu Threshold Label Volume


Done:

  • fiducial behavior
    • tractography updates not happening during drag - fixes checked into SlicerVTK and Slicer3 branch
    • widget issues (Nicole putting fixes in SlicerVTK)
  • making sure mouse-modes work same in 3D/2D (Wendy and Nicole) close!
    • done for 2d, Andras has tested it and says it's fine
  • editor volume merge feature clarification (Steve meeting Sonia June 2, fixes committed June 3)
  • getting BRAINS tests to pass (Hans is working on it)
  • tracking down what is making that crash on Mac OsX when several PET volumes are loaded into Slicer (Wendy).
  • are there extra pick operations going on? There seems to be a delay after clicking mouse before camera begins to rotate when large models are present in scene. could not replicate - tried 2.5 million triangle model on slicer and paraview and performance is the same
  • should picking of fiducials in 3D be disabled unless user explicitly chooses 'pick mode'? Not a good idea to change behavior at this point
    • should parallel behavior be implemented in 2D for consistency? Stevie votes no, Wendy votes yes. Not a good idea to change behavior at this point
  • roundtrip to xnat desktop --OK: confirmed that Ron was using dirty files in cache; checking 'always re-download' in the Cache and DataI/O Manager GUI caused fresh download and successful load into slicer. Sometimes, even when you delete cached files thru the GUI, some applicaiton is holding on to files still there and they don't get cleared out. I'm adding a little code to check for remaining files after a ClearCache() is called; if so, we'll warn user that not all files were removed, and to select 'always re-download' to ensure proper data loading.(Wendy)
  • view control gui: (Wendy) this guy isn't always updating anymore. He may be a casualty of multiple view nodes in the scene. He was a casualty but is fixed now!

Issues that cannot be fixed (or aren't crucial to fix) should be listed in the slicer 3.6 release notes.

  • finishing up the DICOM file copy on remote upload, also close (Wendy)