New remote branch pushed: bug-14882-RepairTutorialStep3

This bug was fixed inclusively by achieving an intermediate objective for T14443.

This bug is fixed as soon as T14443 is merged.

[7d446a]: Merge branch 'bug-14465-movePropertyListXmlIoToMitkExt'

Merged commits:

2013-02-07 18:18:35 Joseph Görres [82d184]
Added four files and added them to files.cmake

New remote branch pushed: bug-14465-movePropertyListXmlIoToMitkExt

New remote branch pushed: bug-14443-MakeImageAccessThreadSafe

[ea44ab]: Merge branch 'bug-14443-MakeImageAccessThreadSafe'

Merged commits:

2013-07-31 17:31:48 Joseph Görres [c220da]
COMP: made sure there is always a valid image data item

[2cd90b]: Merge branch 'bug-14443-MakeImageAccessThreadSafe'

Merged commits:

2013-06-26 16:50:17 Joseph Görres [21bff4]
reactivate multi-threading in ImageAccessorTest

With the ITK4 migration, we finally got rid of the "WEAKPOINTER_PROBLEM_WORKAROUND"

Additionally to the WEAKPOINTER_PROBLEM, which will be fixed with the ITK4 integration, there is another thread problem with the validity of ImageDataItems.
These get reinitialized when making sure, that the given image data is completely initialized. In parallel, this can lead to undefined values. As it can be seen in the ImageAccessorTest, random failes occur when creating ImagePixelReadAccessors concurrently, e.g. ImageDataItem->GetDimension() returns occasionally undefined values.

Main problem is still the "WEAKPOINTER_PROBLEM_WORKAROUND" in mitkBaseData. However, switching off the workaround makes it impossible to use the Image class anymore. This is going to be fixed with the ITK4 migration, and therefore this bug depends on T10925.

Additionally to the typo in the ImageAccessor code snippet, an update of the cloning explanation is necessary.

[8a3be5]: Merge branch 'bug-14356-UpdateImageConceptDocumentation'

Merged commits:

2013-05-08 16:50:26 Joseph Görres [58f9cf]
fixed typo and altered cloning explanation

New remote branch pushed: bug-14356-UpdateImageConceptDocumentation

With the current commit, all mitk::Image::GetData() calls are replaced in mitkImageGenerator.h and mitkImageTest.cpp.

New remote branch pushed: bug-14306-ReplaceDeprecatedMethodCallsInCoreTests

I am working on a more general solution to remove side effects regarding image data access.

Thank you for the detailed bug description and the provided fix. Your fix is good and it is possible to change the ImageDataItem smartpointer to a standard pointer because the ImageDataItem is still hold by a smartpointer in the Image class. However, I would call this fix a workaround.

New remote branch pushed: bug-13953-4DdataCroppingCrash-RemoveSideEffectsInImageDataAccess

[64d68e]: Merge branch 'bug-13953-4DdataCroppingCrash-RemoveSideEffectsInImageDa

Merged commits:

2013-01-16 14:55:03 Joseph Görres [066782]
replaced Image smartpointer to a weakpointer in ImageVtkAccessor

The bug is fixed with the current bug branch.

The test doesn't fail anymore. => resolved

New remote branch pushed: bug-13935-ImageAccessorTestNew

New remote branch pushed: bug-13935-ImageAccessorTest

Most methods in mitk::Image aren't thread-safe. This test makes sure, that the recently introduced Image accessors work correctly to provide at least a thread-safe image access.

With the current master, the application doesn't seem to crash anymore. However, the Application stops after showing the four window view. There are no other GUI items visible (no Datamanager or Menubar). The Application also doesn't response to anything but closing its console view, leading to a proper shut down of MITK.

New remote branch pushed: bug-13526-IntegrateModifyIntoImageAccessor

1. I asked him in person and his particular case was covered. But you’re right, in general it would be better to implement a mitk::Image::Modified() method. But it's not that easy (see next remark)

Since an image can have several instances of vtkImageData, it is not possible to address the right vtkImageData on a global Modified()-call for the whole image. Therefore, we implemented the calls into the destructor of ImagePixelWriteAccessor, which has access to a given ImageDataItem and its vtkImageData.

[b37e26]: Merge branch 'bug-13410-AddFlipImageFilterToBasicImageProcessing'

Merged commits:

2012-10-11 20:15:50 Joseph Grres [864385]
Added FlipImageFilter to BasicImageProcessing with one parameter for the axis

We further investigated that behavior by debugging CastToMitkImage step by step. No errors were detected. The strange behavior was due to wrong usage of CastToMitkImage in mitkToFMpiPointNormalsFilterTest in T13322.

It wasn't necessary to replace ExtractImageFilter with ExtractSliceFilter.

[c2493a]: Merge branch 'bug-13309-OverwriteSliceImageFilterTest'

Merged commits:

2012-10-24 14:27:23 Joseph Grres [0dcea1]
Reactivated test and repaired OverwriteSliceImageFilterTest

The deadlock occurs in CompareImageSliceHelper::CompareSlice, which is called in line 111. A separate slice image created from ExtractImageFilter is compared with a slice of its 3D input image. However, CompareSlice tries to compare the 3D input image with itself. This behavior does not make any sense, is not intended and leads to a double access request from the same thread and thus to a deadlock.

2D and 4D Tests seem to work properly. Deadlock occurs only in 3D Test.

Unfortunately the SegmentationInterpolationTest depends on the OverwriteSliceImageFilterTest. Thus, it was necessary to disable that test too.

Support for Windows XP has been recovered with branch bug-13230-ImageAccessorIntegrationThreadIdIntegration. Release-Flag necessary for merging!

[07df08]: Merge branch 'bug-13230-ImageAccessorIntegrationThreadIdIntegration'

Merged commits:

2013-01-30 16:04:42 Joseph Görres [006a94]
Improved defs in ImageAccessorBase and changed to exception macro in ImageAccessorTest

New remote branch pushed: bug-13230-ImageAccessorIntegrationThreadIdIntegration

[f91aee]: Merge branch 'bug-13230-ImageAccessorIntegrationThreadIdIntegration'

Merged commits:

2013-03-13 17:39:00 Joseph Görres [287023]
Merge remote-tracking branch 'origin/master' into bug-13230-ImageAccessorIntegrationThreadIdIntegration

[d3e84d]: Merge branch 'bug-13230-ImageAccessorIntegration'

Merged commits:

2012-09-27 18:27:17 Joseph Grres [f09e67]
Made former image access methods deprecated and integrated image accessors implicitly in itk cast- and access-methods.

[fb21f3]: Merge branch 'bug-13230-ImageAccessorIntegration'

Merged commits:

2012-10-01 10:02:58 Joseph Grres [c39326]
COMP: Disabled also mitkSegmentationInterpolationTest temporary (See T13309)

[2110d1]: Merge branch 'bug-13230-ImageAccessorDocumentation'

Merged commits:

2012-10-31 14:19:36 Joseph Grres [4659a9]
Example source code in concept page was not compilable

[2cae8e]: Merge branch 'bug-13230-ImageAccessorIntegration'

Merged commits:

2012-09-28 14:54:34 Joseph Grres [c33b23]
COMP: Dealt with compiler specific errors and repaired some class name conflicts

[282572]: Merge branch 'bug-13230-ImageAccessorIntegration'

Merged commits:

2012-09-28 18:33:12 Joseph Grres [15b27f]
COMP: Disabled OverwriteSliceImageFilterTest temporary: See T13309

[1167c4]: Merge branch 'bug-13230-ImageAccessorIntegration'

Merged commits:

2012-09-28 18:09:45 Joseph Grres [63b786]
COMP: Repaired broken ImageAccessorTest

[32e796]: Merge branch 'bug-13230-ImageAccessorIntegration'

Merged commits:

2012-10-01 10:20:14 Joseph Grres [f911ce]
COMP: Prevent SegmentationInterpolationTest from execution

[00e4ed]: Merge branch 'bug-13230-ImageAccessorIntegrationFixes'

Merged commits:

2012-11-07 18:00:54 Joseph Grres [68a8e4]
Simplified redundant code and changed pixel type check to allow the use of compound pixel types

[42a4ff]: Merge branch 'bug-13230-ImageAccessorIntegration'

Merged commits:

2012-10-01 13:36:22 Joseph Grres [036ee5]
Added deprecatedSince tags to Doxygen and fixed some whitespace style issues

There seems to be no implementation for non-closed planar figures in the ImageStatisticsCalculator. A separate treatment of non-closed planar figures has to be implemented from scratch.

The following error is thrown:

GetPixelValueByIndex is deprecated. This problem is solved with mitk::ImagePixelReadAccessor::GetPixelByIndexSafe.

If all exisiting position nodes are removed before scene saving, all segmentation variations (3d, 2d and disabled) can be saved and loaded properly. That means, if mode show positions is activated, segmentations cannot be saved.

screenshot of kcachegrind

Valgrind callgrind showed, that 60 % of the computation time is spent in a method called GetScalarValueMin(int).

the whole callgrind graph

In Linux loading and saving time vary heavily. Sometimes 8 images can be loaded immediately (1-2 seconds), sometimes the loading process is more than 10 times slower...

the whole callgrind graph

screenshot of kcachegrind

Does this error occur for "single screenshot" too? A short look into the code showed that the option "multiplanar screenshot" depends on the QmitkStdMultiWidgetEditor. However, the single screenshot option directly retrieves the VtkRenderer from current focus. The retrieved VtkRenderer should be independent of the QmitkStdMultiWidgetEditor.

[44d389]: Merge branch 'bug-11569-RemoveIIL'

Merged commits:

2012-04-18 09:38:40 Thomas Kilgus [1a74ee]
Added X11 config for MITK and dependency from GPGPU to X11.

No it is not. Due to a missing const-correctness in Image and ImageDataItem, ImageAccessors also need a non-const pointer. One reason for this problem is T13324

The brain.mhd file has a transformation matrix, which should reflect the image across the x-axis. However, if the diagonal elements of the transformation matrix are negative, the bug will occur.

This bug is really disturbing when storing many items. If you want to show images/objects to a physician, it is helpful to create a big project file with many items, which then can be assessed very fast item by item without interrupts like loading or searching files. In this case a manual reorganisation of all data nodes can easily take more than 10 minutes, especially if an alphabetical order is desired.

[70f75c]: Merge branch 'bug-11096-ReworkTutorialCode'

Merged commits:

2012-04-13 13:40:32 Joseph Grres [486723]
Step9 updated

New remote branch pushed: bug-10925-ITK4Segmentation2DInterpolationFix

Current condition: CMake Configure works, but Generate throws following error:

This bug has been resolved within T13410.

[5c33a9]: Merge branch 'bug-1-umlaut-fix2'

Merged commits:

2013-01-23 15:57:25 Joseph Görres [f4951e]
Test

New remote branch pushed: bug-1-umlaute-test

[aef185]: Merge branch 'bug-1-umlaute-test'

Merged commits:

2013-01-23 15:43:52 Joseph Görres [d7360d]
new file