This bug seems to be indeed a problem with loading 4D images. It can be reproduced by loading the US4DCyl.nrrd from MITK-Data, saving it to a different file location and loading it again. The difference between the source image and the stored one is that the stored one is zipped.
Aug 2 2016
Bug has been resolved within T12953. Saving of 4D images works now, hence the saving of scene files with 4d images works as well.
Scenes with 4D data provoke a crash of the MITK workbench.
In a first step this crash needs to fixed.
As a less critical feature a method for handling 4D data in scene files needs to be implemented.
Loading of scene files containing 4D images does not work.
I checked the affected plugins in more detail.
The error occurs using the following plugins:
I cannot reproduce this bug with newer builds of MITK. Probably this bug was accidentially fixed by an other bug.
[fbbe42]: Merge branch 'bug-16527-RemoveQmitkSegmentationPostProcessing'
2014-01-15 14:17:28 Sven Mersmann [f5882e]
Remove QmitkSegmentationPostProcessing from repository as it is no longer needed.
This problem occurs on a CTK version that is currently not in use in MITK. Please decide whether it is release relevant once CTK gets updated.
Implemented the unit test and integrated it in the testing framework. Set Needs_Core_Modification and release_fix flags.
[f9501b]: Merge branch 'bug-16300-EnhanceCompareFilterWithThreeComponentImages'
2013-11-20 15:17:44 Sven Mersmann [2bdb50]
COMP: Added missing include.
Rename the mitkCompareImageFilter to mitkCompareImageDataFilter since the data are compared and not the entire image.
Merged bug branch into master
Once the mitkCompareFilter works with three component images, the Equal() method needs to be adapted and integrated into the mitkOpenCVToMitkConversionTest.
New remote branch pushed: bug-16300-EnhanceCompareFilterWithThreeComponentImages
The image comparison for equality does crrently not work with three component images...but it should.
Included the filter in the equal method and tested it with mitkOpenCVToMitkImageConversionFilter, which worked. Now I write a unit test for the class.
[a07de3]: Merge branch 'bug-16300-EnhanceCompareFilterWithThreeComponentImages'
2013-11-13 16:49:33 Sven Mersmann [a5093c]
Integration of mitkMultiComponentImageDataComparisionTest as CUSTOM_MODULE_TEST.
[e9d2ef]: Merge branch 'bug-16300-EnhanceCompareFilterWithThreeComponentImages'
2013-11-20 15:02:16 Sven Mersmann [8de253]
COMP: Removed warnings and fixed the test.
[c27890]: Merge branch 'bug-16300-EnhanceCompareFilterWithThreeComponentImages'
2013-11-20 16:40:15 Sven Mersmann [2b404a]
COMP: Added passing of filter result to output.
New remote branch pushed: bug-16278-ExtendTestForOpenCvImageConversion
[d23d48]: Merge branch 'bug-16278-ExtendTestForOpenCvImageConversion'
2013-10-21 13:23:32 Sven Mersmann [0a873d]
Enhanced the unit test for conversion of mitkImage to OpenCV image and back. Excluded the itkImageBridge for conversion to mitk for floating point images do not work with this class. Included namespace mitk.
[1cbbd8]: Merge branch 'bug-16278-ExtendTestForOpenCvImageConversion'
2013-10-21 14:57:55 Sven Mersmann [b9ffac]
COMP: Changed release of OpenCV cv::Mat object, which is not a pointer.
The image conversion should now workfor all kinds of images. To test the three component images the mitkCompareFilter and accordingly the Equal() method needs to be modified (T16300).
New remote branch pushed: bug-16254-ImplementSetLargestPossibleRegion
[4930d6]: Merge branch 'bug-16254-ImplementSetLargestPossibleRegion'
2013-10-11 14:14:33 Sven Mersmann [2825d2]
Method to set the largest possible region in a sliced data object implemented.
Duplicate of T16223
Fixed within T13180. Bug is closed.
This bug seems to be a more general case and related to the T13180 (The progress bar is calling qApp->processEvents that results in enabling user interaction while synchronous process is going on).
It would be better to solve the main issue, and in this regards close this bug and set the bug-13180 to higher importance.
It seems like this bug is related to the plugin list created by the ctkPlugin framework. At a certain point, which is not clearly reproducable yet, the plugin list is corrupted and the execution crashes, but only after /BlueBerry.clean.
I'll check this issue!
I think the test still uses the old test macros. These should be replaced by the new ones.
The mitkWorkbench crashes while closing the application due to uninitialzed variable "plugin" in ctkPluginContextPrivate (l.42).
New remote branch pushed: bug-15969-EditableTooltip
Some unit tests are failing due to the conversion of double/float in the entire MITK.
These tests are:
Results pushed to bug-15963-IGTCodingBase2.
[1f3cf6]: Merge branch 'bug-15741-NewPickingOption'
2013-09-05 15:43:34 Sven Mersmann [bf4fdc]
Enhanced the picking option by CellPicking. The documentation gives some hints about usage of the different pickers.
Changed hardware flag to linux according to Michaels message.
New remote branch pushed: bug-15741-EnhanceVtkPropRenderer
This bug has been discussed in he MITK meeting and will be solved as described above.
New remote branch pushed: bug-15741-NewPickingOption
Yes, No, Yes, in this order!
Their is another picker called vtkPropPicker, but I don't know if we really need this one.
The best docmentation on the picker is in the vtk book, so we could refer to this for detailed documentation.
All occurences of GetData() in ToF moudule removed. Bug is fixed.
[9d9cd1]: Merge branch 'bug-15575-RemoveImageGetDataInToF'
2013-07-10 16:13:16 Sven Mersmann [54ea11]
Removed the calls to GetData by calling an instance of the ImageReadAccessor.
[d25103]: Merge branch 'bug-15575-RemoveImageGetDataInToF'
2013-07-10 17:16:57 Sven Mersmann [5d6d28]
COMP: Changed check for empty image, since direct access to the empty data is not possible.
New remote branch pushed: bug-15575-RemoveImageGetDataInToF
New remote branch pushed: bug-15514-RemoveDeprecatedCallsInMitkExt
[e5ba71]: Merge branch 'bug-15514-RemoveDeprecatedCallsInMitkExt'
2013-07-03 17:58:47 Sven Mersmann [9a6f1b]
Removed Image::GetData() method and replaced it with mitk::ImageReadAccessors.
Bug is fixed.
Remove deprecated calls to Image::GetData() in the measurement module and view.
New remote branch pushed: bug-15501-RemoveImageGetDataFromImageTest
New remote branch pushed: bug-15501-RemoveImageGetData()Call
Bug is fixed.
[0cc6d6]: Merge branch 'bug-15501-RemoveImageGetDataFromImageTest'
2013-07-03 15:27:41 Sven Mersmann [dab571]
Removed all direct accesses to the image data and replaced it with mitk::ImageReadAccessors.
Pushed branch with changed mitkImageTest.
One further question, why the FileDialogPattern in mitk::ImageWriter still contains *.pic and *.pic.gz, those formats are deprecated since a while and the write support was stopped already.
Now saving of images is not possible at all, I'll doublecheck this new problem and report to this bug.
First, the motivation of this bug remains unclear, what is the point to save base data?
Is this an flexible extension checking for different data types? Then T8209 should be related to this bug, since this will automatically solve it.
New remote branch pushed: bug-15480-SavingOfImagesDoesntWork
Thanks for the context, I'll check this issue on my windows maschine and try to get a viable solution.
[d523a1]: Merge branch 'bug-15392-RemoveUninitializedVariable'
2013-06-17 10:09:25 Sven Mersmann [b0ce99]
Extended one comment and make it a debug info.
New remote branch pushed: bug-15392-RemoveUninitializedVariable
The fix is merged and works. Bug resolved.
In T13162 a similar problem is described. That can be removed from that bug.
New remote branch pushed: bug-15359-RemoveMisleadingErrors
[8650c1]: Merge branch 'bug-15359-RemoveMisleadingErrors'
2013-06-19 17:48:42 Sven Mersmann 
Remove misleading error messages. For open planar figures it is currently intended behavior to not show the statistics. Refer in this context also to T13162.
After doublechecking the issue it is more a usability problem. When I generate a new segmentation, I expect to be able to immediatly work with segmentation. It should hence be considered to automatically select the new segmentation as the one that is manipulated.
I investigated this issue today using Jan's expanded unit test. It seems like the memory management is corrupted. Whit using mitk::ItkImageImport the memory is managed as a reference of the filter. So the image did not allocate its own memory field. Once the methods gets out of scope, the reference on the filter is lost and so are the data.
This problem can be solved by using the method mitk::GrabItkImageMemory(...) which is also implemented in the mitkITKImageImport class.
This method takes care of the allocated memory regardless of the output from the threshold filter.
This issue seems to be fixed?
Andi, did you work on this issue? There is nothing described in this bug!
Eric, if you work on this bug, please set it to IN_PROGRESS to prevent that somebody else takes it.
New remote branch pushed: bug-14930-LiverSurgeryPlanning_extern
New remote branch pushed: bug-14930-MITKLiverSurgeryPlanning_rebase20130902
[85d649]: Merge branch 'bug-14736-ExecutionOfRigidRegistration'
2013-07-31 16:51:46 Sven Mersmann 
Separate the image data of fixed and moving image from the view to prevent double memory access, which results in a thrown exception.
Pushed to master, solution seems to work. Bug is fixed