Is this still a valid issue?
- Queries
- All Stories
- Search
- Advanced Search
- Transactions
- Transaction Logs
All Stories
Jan 20 2022
Jan 19 2022
Was adapted for the 2021.10 release with the following findings:
In T28617#226314, @kalali wrote:Simply removing this function does not work since it's the only function that allows to pass a specific single render window. And it will be used more often when using the mxn-multi widget.
However, it is not complicated to refactor these functions to remove this parameter and to simplify the functions to call each-other to allow passing a single render window.
Jan 18 2022
Was adapted for the 2021.10 release with the following findings:
- T28826: [Segmentation][Interpolation] Color of 2D interpolated slice is not yellow but segmentation color
- T28827: [Segmentation][Interpolation] Color of 3D interpolated segmentations is red, not green
- T28829: [Segmentation][Interpolation] Interpolation area is still enabled, when segmentation node is hidden.
Should be taken care of with T27873: Extend "Checkliste Segmentierung – Interpolation" when T28142: [Segmentation] Remove plugin redundancy with MultiLabelSegmentation is done.
Was adapted for the 2021.10 release with the following findings:
Was adapted for the 2021.10 release with the following findings (more related tasks see comment before):
- T28798: [Segmentation][MultiLabel][Utilities] UI is inconsistent / unstable
- T28801: [MultiLabel Segmentation][Utilities] Exception thrown when converting to MultiLabel
- T28796: [Segmentation][Utilities] Morphological Operations behave unexpectedly
- T28797: [Segmentation[MultiLabel][Utilities] Using 4D data with dynamic segmentations results in 3D data
@s349i Maybe you can say something about the expected output / generated segmentation. Also a reminder for the upcoming Segmentation-discussion.
Was adapted for the 2021.10 release with the following findings:
- T26959: [Segmentation] Newly created LabelSet segmentation is not displayed as "outline binary"
- T28766: [Segmentation] Preferences / properties for "Smoothed surface creation" do not seem to have any effect
- T28789: [Segmentation][Multilabel][Options] Auto-selection mode is unclear in checklist or broken
- T28791: [Segmentation][Multilabel][Options]"Create polygon model" issues for large images
- T28793: [Segmentation][Multilabel][Options] Changing only decimation rate has no visible effect in "Create smoothed polygon model"
Was adapted for the 2021.10 release with the following findings:
- T26959: [Segmentation] Newly created LabelSet segmentation is not displayed as "outline binary"
- T28766: [Segmentation] Preferences / properties for "Smoothed surface creation" do not seem to have any effect
- T28789: [Segmentation][Multilabel][Options] Auto-selection mode is unclear in checklist or broken
- T28791: [Segmentation][Multilabel][Options]"Create polygon model" issues for large images
- T28793: [Segmentation][Multilabel][Options] Changing only decimation rate has no visible effect in "Create smoothed polygon model"
It's not a duplicate but the bug is known and probably related to T26959: [Segmentation] Newly created LabelSet segmentation is not displayed as "outline binary".
In T26959#220748, @kislinsk wrote:We are going to rewrite the 2-d LabelSetImage mapper and write a 3-d LabelSetImage mapper this year, where this will be addressed. We should at least clean up the duality of the contour properties (binary vs. labelset.contour). We should probably allow to show the contour without any filling and we should clean up the coupling between contours and active label.
Jan 17 2022
Jan 15 2022
Jan 13 2022
Jan 12 2022
The current solution was never meant as a user interface but only for saving slice positions in scene files. However, it turned out that it is quite handy to jump to drawn contours with a simple click so we should probably move the list of positions into the segmentation view itself if needed.
Suggestion: Open another task for ensuring at least default properties like "visible" that we can then kind of rely on.
We can assume that this is a time-intensive task and a few solutions were already proposed but we do not see how to resolve it completely at the moment with our available resources.
Jan 11 2022
In T28903#232375, @kalali wrote:However, I agree that it should be enough to only care for the validation of the currently selected input nodes (working node / reference node).
I can propose a change and we can test if this is the desired behavior.
That would be nice. Thanks.
Good idea to talk about it in the group meeting.
We should discuss if it makes sense to make these warnings consistent if we eventually merge both Utilities views into one (as mentioned by @s349i).
Maybe I can clarify something:
Thank your for the report. I removed all erroneous MITK_DEBUG calls as they should be used rarely anyway.
Deleted branch from rMITK MITK: bugfix/T28910-MITK_DEBUG-Errors.
Pushed new branch to rMITK MITK: bugfix/T28910-MITK_DEBUG-Errors.
Since @floca suggested we merge Segmentation Utilities and MultilabelSegmentation Utilities only one of each Widget will probably remain. But also since we only want one of the SegmentationUtilities and MultilabelSegmentationUtilities probably the mitkMultiLabelImage class should be integrated now since I am still using the LabelSetImage as a makeshift. I have made some progress on BooleanOperations and MorphologicalOperations. Probably something to talk in the coming group meeting.
In T28176#229224, @s349i wrote:Testing if both Segmentation Utilities and Multilabel Segmentation Utilities have the same functionality:
...
@s349i Are you aware of this task? What is your current status conc. the two Utilities Views?
The changes have already been merged. If there are no critical bugs found since the last release (where the changes of the related D533 were already included) we can continue with this task / subtasks.
I'd like to discuss our goals concerning this task for MITK v2022.04.