For both bugs, small level window ranges are problematic. Dependency can be upheld until there is further clarity on 14652.

[0cbc70]: Merge branch 'bug-19193-PointSetNotDisplayedAfterReading'

Merged commits:

2015-08-12 14:37:39 Tobias Norajitra [ab9e83]
Proper model reset for point list widget update.

Erroneous point list model reset calls in QmitkPointListModel::SetPointSetNode() led to the inproper update behaviour of the point list widget. Bug has been fixed by proper adjustment of point list model reset.

[6937b5]: Merge branch 'bug-19180-PointSet-ManuallyAddPointBug'

Merged commits:

2015-08-05 14:30:21 Tobias Norajitra [fac6ba]
Use max id convenience method instead of faulty size id.

The reason for the bug lies within mitk::PointListWidget::OnBtnAddPointManually(), where the max point id is being determined using mitk::PointSet::GetSize(). If a non-max id point has been deleted beforehand, a false max id is delivered, and the actual max-id point is modified instead of a new point insertion.

Bug is solved using mitk::PointSet::GetMaxId() where proper max id is delivered regardless of prior point deletion.

Currently, unique imageForLevelWindow selection has been fixed for 1. manual selection via context menu of the level window slider, 2. for identical images (see T19436) and 3. for node visibility changes. Still, unique manual selection via property list is not supported yet.

An new test case has been introduced in the level window manager test to validate random manual settings of the imageForLevelWindow property.

Unique manual setting of the imageForLevelWindow property has been added by adding the level window manager as an observer to the property. Consequently, the creation of the property has to be made upon observer list update. In addition, data node deletion is now taken into account.

[0dc11e]: Merge branch 'bug-19150-integration-0'

Merged commits:

2016-03-16 17:03:38 Tobias Norajitra [b16bdb]
Unnecessary code lines removed.

[9db8f0]: Merge branch 'bug-19150-integration-0'

Merged commits:

2016-06-01 16:15:33 Tobias Norajitra [e0224f]
COMP: unused variable removed to avoid warning.

Test case added to the level window manager test, where the imageForLevelWindow property is checked against random data node visibility changes.

The erroneous observer removal is caused by a missing deletion of the m_PropertyDeletedTags map in QmitkPropertyItemModel::SetNewPropertyList. An appropriate deletion fixed the issue. According test cases have to be added next.

According cases have been added to the Update-Function of the LevelWindowManager. However, another bug is preventing proper updates on visibility changes. For each node, an observer is added to the visible-property by the level window manager. This observer is removed on data manager selection changes if the property list view is opened.

In fact, the if-selection at the end of the LevelWindowManager::Update method seems to be senseless and has been removed without apparent effect.

The bug originates from the LevelWindowManager::SelectLevelWindowProperty method. There, for the image that is associated with the handed over levelWindowProperty, the imageForLevelWindow is correctly set. Still, the image holding the former m_LevelWindowProperty keeps it imageForLevelWindow unchanged. Issue is fixed by deactivating imageForLevelWindow for the image associated with m_LevelWindowProperty prior to setting the new levelWindowProperty.

[e61d51]: Merge branch 'bug-19145-ULThresholdSilder-IntegerPixelTypeHandling'

Merged commits:

2015-08-19 16:48:49 Tobias Norajitra [36a440]
Slider interval settings adapted for non-float values and proper input value calculation.

In order to fix this, the slider value types are adjusted to the input image scalar type. Thus, the user is prevented from setting non-integer values for integer scalars.

[cbee9b]: Merge branch 'bug-19138-PointSetLacksConvenientPointRemoval-int0'

Merged commits:

2015-07-15 16:32:56 Tobias Norajitra [da13a5]
Point removal testing cases added to mitkPointSetTestSuite.

For testing purposes, mitkPointSetTestSuite has been extended by the according test case. That is, for the default point set, remove point is called for non-existent points and non-existent time steps, assuring no removal is performed in these cases. Point removal of max id and non-max-id points is performed and compared to accordingly provided reference point sets.

The point set API has been extended by convenience methods for a) a removal of a point from the point set if that particular point exists and b) a removal of a max id point from the point set.

There is no obvious reason why the threshold access methods had to be hidden. The according changes have already been reverted.

[7834b4]: Merge branch 'bug-19013-RegionGrowingSliderAdjustment'

Merged commits:

2015-10-21 16:03:43 Tobias Norajitra [91e27c]
Max value should be based on the seedpointvalue rather than on the lower threshold.

Currently, after region growing, the maximum slider value is calculated based on the lower threshold. If for a fixed upper threshold the lower threshold is decreased, the maximum slider value is decreased consequently, while the maximum value of the region grower output remains constant. Therefore, the maximum slider value should remain unaltered.

Calculating the slider max value based on the seed point value should fix the issue.

This bug could not be reproduced on behalf of T14605 and seems having been fixed there.

[aa49ea]: Merge branch 'bug-18984-3DRegionGrowing-sliderAdaptiveVolumeRendering'

Merged commits:

2015-10-21 17:01:00 Tobias Norajitra [a15860]
COMP: invalid typenames removed.

This bug has already been reported for bugs 13162 and 14605. Solved on behalf of 14605.

[9453f7]: Merge branch 'bug-18984-3DRegionGrowing-sliderAdaptiveVolumeRendering'

Merged commits:

2015-10-21 16:31:32 Tobias Norajitra [86e6ec]
COMP: typename added.

[1c2cdc]: Merge branch 'bug-18984-3DRegionGrowing-sliderAdaptiveVolumeRendering'

Merged commits:

2015-10-21 15:59:50 Tobias Norajitra [53f101]
Update masking threshold on volume rendering activation.

[33aa17]: Merge branch 'bug-18984-3DRegionGrowing-sliderAdaptiveVolumeRendering'

Merged commits:

2015-10-21 16:16:11 Tobias Norajitra [1a42a8]
COMP: typename added.

This might be due to the write access on data that is being read by the interpolator. The crash appears to be more likely to be reproduced if the interpolation is more complex which seems to underline that assumption. Interestingly, if you perform the undo / redo operations once in sufficiently slow speed, it is more difficult to evoke the crash afterwards.

It has been conferred that the tbss view should be removed for the upcoming release as that functionality is being provided using FSL by default. Until further demand.

[c43aa8]: Merge branch 'bug-18278-OtsuSegmentationPreviewInactiveOnCancelingDial

Merged commits:

2014-10-15 16:21:30 Tobias Norajitra [cf2edf]
Open the warning dialog prior to number of regions actualization, depending on the region number spin box value instead.

Sorry, missed the detail.

Duplicate bug of 18278 which has been solved.

The problem is located in the OnBinaryPropertyChanged() method, which is called only upon unchecking the segmentation binary property. The patient image uninentionally goes to the segmentation image selector due to its binary property. Then, the function returns prior to consideration of the segmentation image binary property swap.

Can be solved by reinit on the segmentation and will be obsolete in the near future.

The crash occurs because on visibility change, no data is selected in the segmentation image selector. For that case, a separate tool selection box check solves the issue.

Moving to segmentation component as only QmitkSegmentationView is involved.

[7a6829]: Merge branch 'bug-17901-CrashInRegionGrowingPlugin'

Merged commits:

2014-09-24 16:51:32 Tobias Norajitra [ca5c5b]
Replace ImportItkImage by CastToMitkImage to allow out-of-scope memory management.

Bug is solved. Closing.

After region growing calculation, the result gets lost after conversion to mitk image when leaving scope. This is due to the remote memory management of ImportItkImage. Using CastToMitkImage instead solves the issue.

[5a8670]: Merge branch 'bug-16698-FixPointRemovalInPointSetDataInteractor'

Merged commits:

2014-08-20 14:25:31 Tobias Norajitra [c51c68]
Comments removed.

[a1597a]: Merge branch 'bug-16698-FixPointRemovalInPointSetDataInteractor'

Merged commits:

2014-09-17 14:51:20 Tobias Norajitra [7243f3]
mitk-data revision tag adjusted.

Everything seems to work, so we can close this bug.

Thanks for your fix, it seems to work fine for both the PointSetInteractor and the PointSetDataInteractor. I will push with the cherry-picked fix then.

(In reply to Miklos Espak from comment #7)

Have you got the core change request affirmation?

Hey Miklos,

We are waiting for the core change request affirmation and see into merging the fix asap.

This bug is indeed partially related to T15264 where according changes have been made with regard to the misuse-case within the point set interactor. The vtk mapper case should be resolved as stated. The else-case seems to be non-problematic as iterators are used to traverse the points container.

The interpolation deactivation on view change is regarded secondary here. Apart from that, the issue is solved.

[4aebeb]: Merge branch 'bug-16461-integrationBranch'

Merged commits:

2013-11-20 18:59:31 Tobias Norajitra [88123f]
Preview and segmentation output value for in-region-voxels set to 1. Added entries to QmitkSlicesInterpolator::OnInterpolationMethodChanged() in order to disable 2d slice interpolation in the proper cases.

New remote branch pushed: bug-16461-MorphologicalOperatorsNotFunctioning

Can be reproduced as followed:

1. Open an image, e.g. pic3D.nrrd
2. Create a new segmentation within the segmentation plug-in.

2a. Use the 2D-slice interpolation for manually creating a 3D segmentation.
3a. Apply the Dilation or Fill Hole Operator to the newly created segmentation -> Crash.

Reason for the morphological operator on otsu segmentation fail is that within the produced otsu segmentation, in-region-voxels do not have value 1 (but 255). The segmentation produced from the respective preview image is a result of itk::BinaryThresholdImageFilter application. Output in-region-voxel value has been set to 1 now.

For the slice interpolation case, the crash occurs as the segmentation interpolation controller still observes changes on the segmentation after the segmentation view has been left. This can now be prevented by explicitly disabling the 2D slice interpolation ( according entries have been inserted in QmitkSlicesInterpolator::OnInterpolationMethodChanged() ). Nevertheless, automatically disabling slice interpolation on view change still has to be adressed.

New remote branch pushed: bug-16420-VigraTestIntegration

New remote branch pushed: bug-16420-TestIntegrationOfVigra

[e19b63]: Merge branch 'bug-16382-PointSet-MaxId-InsertPoint-Test'

Merged commits:

2015-07-15 16:37:29 Tobias Norajitra [d9234f]
GetMaxId and InsertPoint testing cases added to mitkPointSetTestSuite.

New remote branch pushed: bug-16382-PointSetInterfaceExtension

[641dae]: Merge branch 'bug-16382-PointSetInterface-int0'

Merged commits:

2015-05-20 17:05:37 Tobias Norajitra [5773b3]
Point set interface extension.

New remote branch pushed: bug-16382-PointSetInterfaceAdaptation

The PointSetInterfaceExtension branch is deprecated.

Seems to be fixed.

[bf92e1]: Merge branch 'bug-16133-3D-interpolation-4D-support'

Merged commits:

2015-04-08 14:31:58 Tobias Norajitra [75eeca]
COMP: check selected segmentation to prevent segmentation tool interaction test fail.

[77b893]: Merge branch 'bug-16133-3D-interpolation-4D-support'

Merged commits:

2015-04-08 13:35:32 Tobias Norajitra [672035]
COMP: Const non-const assignment fix.

[55a7eb]: Merge branch 'bug-16133-3D-interpolation-4D-support'

Merged commits:

2015-04-08 13:53:19 Tobias Norajitra [277a04]
COMP: signed - unsigned assignment fix.

[8240ae]: Merge branch 'bug-16133-3D-interpolation-4D-support'

Merged commits:

2015-04-09 16:34:14 Tobias Norajitra [2963b8]
Merge branch 'bug-18865-SurfaceInterpolationControllerTest-4DsupportExtension' into bug-16133-3D-interpolation-4D-support

[1df760]: Merge branch 'bug-16133-3D-interpolation-4D-support'

Merged commits:

2015-04-01 17:37:59 Tobias Norajitra [1eddc4]
Toggled surface segmentation invisible with regard to 3D+t-data.

Checks for 3-dimensionality had to be adapted as 3d+t data were rejected otherwise. The issue has been solved using time selector output for check.

On time step change, the interpolation pipeline of the SurfaceInterpolationController needs to be reset in order to handle different timesteps independently.

Within the interpolation pipeline in the surfaceInterpolationController, the normal computation filter needs segmentation input for the proper timestep.

mitkSurfaceInterpolationController has been extend in order to hold and organize contours for several timesteps. Furthermore, a time step member has been introduced. Surface interpolation will then be accessed via mitkSegTool2D (e.g. ::UpdateSurfaceInterpolation) and via QmitkSlicesInterpolator. Latter has been adjusted to properly set the newly created timestep member within the mitkSurfaceInterpolationController.

New remote branch pushed: bug-15992-ConditionalLevelWindowSliderDeactivationTest

[9dfc2b]: Merge branch 'bug-15992-CondLvlWinSliderDeactivationTest'

Merged commits:

2013-09-04 19:05:17 Tobias Norajitra [e4c8a5]
mitkLevelWindowManagerTest extended by according test functions.

The mitk::GeometryClipImageFilter delivers wrong results. That is, it internally uses Geometry2D::IsAbove for voxel-wise clip decision where Geometry2D::IsAbove simply compares the voxels' z-world-coordinates against the Geometry2D::BoundingBox min-z-coordinate. In addition, Geometry2D is considered deprecated due to recent re-engineering of the geometry-concept.

Deprecated API calls have been replaced by using proper accessor class instances. Compare with T15640 for further explanations if needed.

New remote branch pushed: bug-15645-DeprecatedMitkImageAPICallsInMitkOverwriteSlicerFilterTest

GetPixelValueByIndex() calls in mitkExtractSliceFilterTest belong to the deprecated part of the mitkImage class. Thus, they have been replaced by mitkPixelReadAccessor instantiations and the corresponding GetPixelByIndex() methods. Changes have successfully been tested on all supported platforms. Waiting for affirmative CoreModification flag in order to push the changes.

New remote branch pushed: bug-15640-DeprecatedMitkImageApiCallsInMitkExtractSliceFilterTest