New remote branch pushed: bug-15428-OtsuDefaultLevelWindow

New remote branch pushed: bug-15420-mitkOtsuSegmentationFilterReimplementation

[b61bf3]: Merge branch 'bug-15368-OtsuSegmentationCrash'

Merged commits:

2013-07-10 16:50:43 Tobias Norajitra [0fbd66]
Replaced itkImageImport in QmitkOtsuAction by GrabItkImageMemory

New remote branch pushed: bug-15368-OtsuSegmentationCrash

Problem was that mitk::ImportItkImage was used for itk-mitk-image conversion. Solved by replacing with mitk::GrabItkImageMemory.

Proper otsu segmented image created before crash.

Proper otsu segmented image created before crash.

No error is thrown during otsu filter execution. The crash occurs for reasons unknown within the vtkImageReslice filter which is invoked after completion of the otsu calculation. The crash does not occur in the redesigned otsu segmentation tool. Further investigation is required.

New remote branch pushed: bug-15318-MultiThresholdOtsuFilterImplementation

New remote branch pushed: bug-15264-PointSetInterfaceChanges

Removing dependency from 16698, as that bug can be solved independently from the point set changes.

FileNames containing misuse cases.

New remote branch pushed: bug-15263-PointSetCloningBug

New remote branch pushed: bug-15264-PointSetMisuseProperInterfaceUse

Topic has been handled by determining the correct number of stored points. This is done by retrieving the index of the last point entry via iterator in the encapsuled itk map container.

bug-15264-PointSetMisuseCorrection and bug-15264-PointSetMisuseProperInterfaceUse are deprecated.

FileNames containing misuse cases.

When points are set in the PointSet, but single indices are left out (as it is the case for index 0 in the example below), the left out points are nevertheless inserted as zero-Points (0,0,0). Consequently, problem arises during cloning within the mitk::PointSet copy c'tor: copying the stored points is performed times the number of only those points that got a valid index before (i.e. not of those left out index zero-Points). Then, there are more stored points than copy actions as the left-out indices do not count for copy, and not all stored points are duplicated.

Separation of segmentation functionality and data representation by integration of mitk::OtsuSegmentationFilter replacing hard-coded itk filter use plus some minor changes.

New remote branch pushed: bug-14929-UnsafeTypeConversionInCellVisitor

The issue is resolved.

Means, already in the master.

[b7a71d]: Merge branch 'bug-14649-TooltipForInterpolation'

Merged commits:

2013-03-20 16:35:39 Tobias Norajitra [2a6d16]
Added tooltips.

New remote branch pushed: bug-14649-TooltipForInterpolation

Image statistics for line profiles are now being calculated within the QmitkHistogramJSWidget based on an according extension of the mitkIntensityProfile. Tabular entries are then generated within a dedicated table view filling method.

[348417]: Merge branch 'bug-14605-StatisticsForLineProfiles-int0'

Merged commits:

2015-06-17 15:58:11 Tobias Norajitra [a7b22f]
Merge branch 'bug-14605-StatisticsForLineProfiles' into bug-14605-StatisticsForLineProfiles-int0

Please add the data involved in bug evocation.

During image statistics calculation, an exception is thrown. This seems to happen in the itkListSampleToHistogramGenerator, supposedly as a consequence of itk::Statistics::Histogram::IncreaseFrequency(). Best guess at the moment is that a static container fails to work on large image data like CTLymphknotenInDerAchsel.pic.gz.

Cannot be reproduced anymore...

As the filename says.

As the filename says.

Erroneous binary image rendering.

Erroneous binary image rendering.

Currently, the binary image rendering does not make mitk crash anymore, so this bug will be closed for now.

Supplement: the level window range has to be adjusted in order to visualize all regions of interest.

Binary image not being rendered properly

Set the proper lut rendering property for multiple threshold otsu segmentation results.

Binary image not being rendered properly

New remote branch pushed: bug-14495-MultipleThresholdOtsuRenderingBug

[cd11b8]: Merge branch 'bug-14495-MultipleThresholdOtsuRenderingBug'

Merged commits:

2013-03-27 15:48:26 Tobias Norajitra [a52af5]
Level window range adjustments

[8cf6c4]: Merge branch 'bug-14495-MultipleThresholdOtsuRenderingBug'

Merged commits:

2013-03-20 16:33:32 Tobias Norajitra [4c0a14]
Set the right rendering property.

Update on the problem: solving T8165 does not neccessarily prevent the zero pointer return. There are rumors that the crash was fixed on behalf of T12056. Still, the binary image rendering shows a very strange and unpredictable behaviour (for an example, see and try attachment).

[6bab0e]: Merge branch 'bug-14483-ImageStatisticsCalculatorVtkSmartPointerIssue'

Merged commits:

2014-02-12 16:23:37 Tobias Norajitra [c4989e]
Image mask member is assigned an image duplicate instead of the importer's output.

The test fail coming along with the proper smart point use goes back to the coupling of the vtkExporter's output bulk data to the internal image mask member of the ImageStatisticsCalculator via itkImporter. That is, after the exporter's deallocation through correct smart pointer reference counting, the member data is being allocated as well by the vtkExporter. The issue has been solved using a persistent image duplicate for the mask member.

(In reply to Tobias Norajitra from comment #5)

Deprecated image API calls replaced.

[1fd8d5]: Merge branch 'bug-14306-mitkImageGeneratorDeprecatedImageAPI'

Merged commits:

2013-07-11 15:44:57 Tobias Norajitra [a9857d]
COMP: type specifier un-typenamed

[98c797]: Merge branch 'bug-14306-mitkImageGeneratorDeprecatedImageAPI'

Merged commits:

2013-07-03 17:56:06 Tobias Norajitra [c97b1b]
Additional Changes.

New remote branch pushed: bug-14306-mitkImageGeneratorDeprecatedImageAPI

Testing implemented as part of bug-15992.

[7e86fc]: Merge branch 'bug-14179-ConditionalLevelWindowSliderDeactivationFix'

Merged commits:

2013-10-09 15:22:42 Tobias Norajitra [2b84a8]
Added an additional property observer to the LevelWindowManager for the rendering mode property, LevelWindowManager::Update being the proper callback function.

[6afc7b]: Merge branch 'bug-14179-ConditionalLevelWindowSliderDeactivationFix'

Merged commits:

2013-10-09 15:49:44 Tobias Norajitra [16c76e]
COMP: Uninitialized variable fix. Zero pointer return handled.

Now, the level window slider is hidden if no image with an applicable level window is present within the data manager.

New remote branch pushed: bug-14179-ConditionalLevelWindowSliderDeactivation

Hiding the level window slider is done by adding a property observer to the LevelWindowManager for the "Image Rendering.Mode" property. The LevelWindowManager::Update function is used for callback, subsequently calling LevelWindowManager::SetAutoTopMostImage. Here, the level window slider is hidden (i.e. m_LevelWindowProperty not being set from NULL), if all relevant data nodes use non-level-window "Image Rendering.Mode" property values.

Could not reproduce the bug. Please provide additional information.

DICOM import button functionality has been reestablished. According implementations concerning file dialog, file selection and dicom import have consequently been added to the DicomView class. Adressing data manager import functionality, ongoing changes to QmitkDicomExternalDataWidget should be considered.

[9d10ac]: Merge branch 'bug-13775-LevelWindowTooltipAdjustmentsMade'

Merged commits:

2013-02-20 15:41:09 Tobias Norajitra [a3a9d3]
Adjustments made

New remote branch pushed: bug-13774-reestablish-custom-viewer-dicom-functionality

Inserted adequate tooltipping text in QmitkSliderLevelWindowWidget for the proper case (the cursor being inside the blue/gray part).

New remote branch pushed: bug-13775-LevelWindowTooltipAdjustmentsMade

Tried to reproduce. Now, the behavior has changed: clicking into the 3D renderwindow once while line tool is activated causes indefinite toolbar deactivation. Soon, the application will crash. To speed things up, just close mitk -> immediate crash.

After some discussion we have come to the conclusion that initializing rendering geometries is an issue that should be handled centrally, and that this bug therefore relies on T12954.

In case of unclear reproducibility: load pic3D.nrrd, right-click on the corresponding datamanager node and choose Otsu-Segmentation from the context menu. In the appearing dialog, select five regions and confirm. Then open the statistics view, select the segmentation data node. The Histogram is shown and does e.g. not show all the values present in the segmentation image.

On windows, it crashes upon DataNode deletion.

This bug is equivalent to T14605 and has been addressed there.

New remote branch pushed: bug-13153-TooltippingAdjustments

For clarification: this bug relates to the Volume Visualization plugin.

[19b84e]: Merge branch 'bug-13153-TooltippingAdjustments'

Merged commits:

2013-02-20 15:04:07 Tobias Norajitra [419cab]
Tooltipping adjustments made.

Actually, all 6 regions of interest are generated. The rendered regions cannot be well differentiated though, as the difference in color is very subtle for two green rendered labels. An image showing the two different green regions is attached. Using another lut for false-color-imaging should be considered.

Subtle color difference while rendering two different labels.

Subtle color difference while rendering two different labels.

Actually, two different problems occur.

New remote branch pushed: bug-13038-noOtsuSegmentationResult

[7939a2]: Merge branch 'bug-13025-LackOfReinitFunctionalityOn2DImages'

Merged commits:

2012-09-26 15:50:01 Tobias Norajitra [b6d4f2]
Changed the interface of OpenRenderWindowPart() to provide programmer's choice for editor activation.

This bug can be seperated into two bugs:

After some discussion we have come to the conclusion that initializing rendering geometries is an issue that should be handled centrally, and that this bug therefore relies on T12954.