Hm, while D480 resolves the issue, I think that the original approach for handling the Multilabel colormap in T23589 is already a workaround. Changing the level window to the maximum possible range is just an emulation of "do not use the level window". We actually have that mode in MITK already. The Image Rendering.Mode property is usually set to LOOKUPTABLE_LEVELWINDOW_COLOR but there is also LOOKUPTABLE_COLOR which is exactly what we need here. The thing is, that LOOKUPTABLE_COLOR is not implemented correctly in the 2-d image mapper, as the level window filter is not reset and instead is using the previously set range. So I will extent D480 to actually use the correct mode and fix the 2-d image mapper.
- Queries
- All Stories
- Search
- Advanced Search
- Transactions
- Transaction Logs
All Stories
Apr 8 2021
If the image is converted to a segmentation (context menu -> Convert to segmentation), it is rendered correctly even though the very same lookup table is used. So I guess the issue is to be found within the 2d image mapper.
The lookup table is set correctly but the rendered colors do not match.
This really shouldn't happen. A first look at the code revealed that the intention is to have at least 25 different colors for values 1-26 and after that repeatedly cycle through 12 colors.
Found the issue. After setting the colormap to "Multilabel" in QmitkDataNodeColorMapAction::OnActionTriggered, the previous LUT is checked if it is a "Multilabel" LUT, instead of the new LUT. Hence the observation above that it works when applying it twice.
This issue was already resolved by @kalali in T23589: Multilabel Segmentations not loaded with multilabel colormap, resp. {D172} for MITK v2018.04.2 but I can confirm that the issue exists for the example image.
Apr 7 2021
I agree with the raised point. I checked the classes again and do not see a valid reason behind the blackbox_func call. I will fix this.
Link to data: https://sas.desy.de/index.php/s/JQgZT5xLCk6RXDg
Does not seem to happen anymore. Can somebody confirm, e.g. @kislinsk?
Sorry, I was on vacation.
Apr 6 2021
We now have some manuals but we still need an example-based guide for reviews/Arcanist.
Apr 2 2021
Apr 1 2021
Mar 31 2021
Created a bug in google gvisor, and one guy is helping us.
Pushed new branch to rMITK MITK: feature/T28406-PatchForTestingLiveWireBehavior.
First patch to change following behavior (for user testing only):
- The linesegments are added with each mouseclick of the users without the cost function but by connecting start and endpoint
- If the user double-clicks on the image, the last point is connected with the first adding an additional segment.