So I am not quite sure that about the re-init problem but to further clarify my list:

I would suggest that you land it for now since it solves the bug and I will create another task in which I also take a look into the stepper problem. But both is fine for me :)

Most likely errors:

• segmentation is always performed on timestep 0 (Threshold/ULThreshold)
• dynamic segmentation is converted into a 3D segmentation (Otsu, Fast Marching, Picking)

Just as a remark: I am getting the exact same error, if I want to load a dcm image from a folder that contains more than one image. Those images originate from a cadaver study which means that before acquisiton most of the patient input data is left blank (study,series etc). If I put the files into separate folders, they are loaded correctly.

I totally forgot to claim the task but I also worked on it *sigh*. Your fix in D346 solves the problem, but I wondered if function initialize() should really be called if the activated() has not been called yet. An alternative would be to have an additional flag "m_activated" or likewise.

Substract Tool also crashes, Paint, Wipe and Correction works

If timestep != 0, the m_CurrentTimeStep is changed with SetCurrentTimeStep(t). In this function Initalize() is called without the m_SmoothFilter initialized before. m_SmoothFilter equals nullptr.

Probably a similar to T24767 (sorry for re-opening but maybe the information from that one helps to narrow it down).
I acutally produced the crash even if the timesteps where valid in both images (contrary to the description).

Although the task was closed I can still reproduce the error:

On the develop (neither debug nor release) I also cannot reproduce the bug, so I think we can close it for now.

But the 3d segmentation has only 1 timestep with duration = sum of durations of all 4D image timesteps, right?

The error described above results from the function add seed that currently takes 3D images only. It can be solved with using a timeselector to crop the current time step and set a 3D seed.
However, the task is related to the discussion on time points versus time steps and should be revised accordingly.

When the workbench is started from scratch it works.

I used a master that was three weeks old, I will double check on the current develop

• Works for dynamic (4D) segmentations but creates the same segmentation for all timesteps
• Could not be reproduced for static segmentations since the workbench crashes during the creation of a static segmentation (related to T27476)

1. Works for timestep 0
2. Independent from segmentation tool so might be a problem of the selection concept not the segmentation

@gaoh: Can you tell what you mean by dynamic mask?

I will leave the task open since the handling in the respective class might need further refactoring to fully understand the behavior (as also mentioned by @kalali in the D279 review).
I remove the MITK2020 tag for now since the mentioned bug is resolved by the fix in D279.

Task T26596 is mentioned but does not correlate with the Topic asaik...maybe transposed digits in the task number?

The reason seems to be that a BoundingShape as such is 3D with an index of size=3 that does not equal the index array of the 4D image. I will test the fix for other images to make sure.

Pushed new branch T27067-ImageCropperCorrect4D.

This was also communicated to Siemens and only occured when loading their early stage clinical prototype data. Can not be reproduced with the newer data. The task can be closed for now.

As stated before, we encountered more examples that raised the unknown exception, so I don’t agree to lower the priority on that.

The latest Release installer