kislinsk renamed T28385: [Segmentation] Erosion crashes the MITK Workbench after interpolation from [Segmentation Utilities] Erosion crashes the MITK Workbench after interpolation to [Segmentation] Erosion crashes the MITK Workbench after interpolation.

Jul 8 2021, 12:23 PM · Cleared, MITK (v2021.10)

kislinsk added a revision to T28606: [Segmentation] Accelerate do/undo mechanism for segmentations: Unknown Object (Differential Revision).

Jul 8 2021, 12:22 PM · MITK (v2021.10)

kislinsk renamed T28606: [Segmentation] Accelerate do/undo mechanism for segmentations from Accelerate do/undo mechanism for segmentations to [Segmentation] Accelerate do/undo mechanism for segmentations.

Jul 8 2021, 12:21 PM · MITK (v2021.10)

kislinsk placed T28480: [Checklist] [Test data] Define publicly available location to store user test data up for grabs.

Jul 8 2021, 12:21 PM · MITK (v2024.06)

kislinsk moved T28614: [Segmentation] Replace ITK filter pipeline for 2-d interpolation with single filter that does the same job from Backlog to Cycle on the MITK (v2021.10) board.

Jul 8 2021, 12:19 PM · MITK (v2021.10)

kislinsk added a revision to T28614: [Segmentation] Replace ITK filter pipeline for 2-d interpolation with single filter that does the same job: Unknown Object (Differential Revision).

Jul 8 2021, 12:17 PM · MITK (v2021.10)

kislinsk added a comment to T28614: [Segmentation] Replace ITK filter pipeline for 2-d interpolation with single filter that does the same job.

Pushed new branch to rMITK MITK: bugfix/T28614-Simplify2DInterpolationFilterPipeline.

Jul 8 2021, 12:16 PM · MITK (v2021.10)

kislinsk triaged T28614: [Segmentation] Replace ITK filter pipeline for 2-d interpolation with single filter that does the same job as Normal priority.

Jul 8 2021, 12:14 PM · MITK (v2021.10)

kislinsk added a revision to T28385: [Segmentation] Erosion crashes the MITK Workbench after interpolation: Unknown Object (Differential Revision).

Jul 8 2021, 7:57 AM · Cleared, MITK (v2021.10)

Jul 7 2021

kislinsk added a comment to T28385: [Segmentation] Erosion crashes the MITK Workbench after interpolation.

Pushed new branch to rMITK MITK: bugfix/T28385-InterpolationCrashOnMorphologicalOperations.

Jul 7 2021, 10:55 PM · Cleared, MITK (v2021.10)

kislinsk added a comment to T28385: [Segmentation] Erosion crashes the MITK Workbench after interpolation.

We can fix it in the SegmentationInterpolationController by explicitly handling the situation that the modified image is not initialized (IsInitialized() == false).

Jul 7 2021, 5:23 PM · Cleared, MITK (v2021.10)

kislinsk claimed T28385: [Segmentation] Erosion crashes the MITK Workbench after interpolation.

Jul 7 2021, 5:14 PM · Cleared, MITK (v2021.10)

kislinsk edited projects for T28385: [Segmentation] Erosion crashes the MITK Workbench after interpolation, added: Cleared; removed Missing Info.

Jul 7 2021, 12:02 PM · Cleared, MITK (v2021.10)

kislinsk added a comment to T28385: [Segmentation] Erosion crashes the MITK Workbench after interpolation.

@floca I just tested it with my latest changes and the issue still persists. The connection to interpolation seems to be that the SegmentationInterpolationController is still holding a reference to the image and is notified about a modification. When changing the segmentation volume to the new segmentation eventually the ImageTimeSelector crashes when trying to clone an ImageDataItem that already seems to be null (this is 0x00000000). Here's a part of the call stack:

Jul 7 2021, 12:00 PM · Cleared, MITK (v2021.10)

Jul 5 2021

kislinsk added a comment to T28606: [Segmentation] Accelerate do/undo mechanism for segmentations.

Pushed new branch to rMITK MITK: bugfix/T28606-UseLZ4InCompressedImageContainerClass.

Jul 5 2021, 11:32 PM · MITK (v2021.10)

kislinsk added a comment to T28606: [Segmentation] Accelerate do/undo mechanism for segmentations.

With the same test data as in T28491, I was already able to virtually eliminate the time spent on compression with LZ4.

Jul 5 2021, 11:03 PM · MITK (v2021.10)

kislinsk moved T28606: [Segmentation] Accelerate do/undo mechanism for segmentations from Backlog to Segmentation on the MITK (v2021.10) board.

Jul 5 2021, 10:51 PM · MITK (v2021.10)

kislinsk triaged T28606: [Segmentation] Accelerate do/undo mechanism for segmentations as Normal priority.

Jul 5 2021, 10:51 PM · MITK (v2021.10)

Jul 2 2021

kislinsk closed T28602: 2021 Week 26 (Very Early July) as Resolved.

Jul 2 2021, 10:50 AM · MITK

kislinsk edited the content of Changelog.

Jul 2 2021, 10:49 AM · MITK

kislinsk added a comment to T28602: 2021 Week 26 (Very Early July).

Deleted branch from rMITK MITK: release/T28602-2021-Week-26.

Jul 2 2021, 10:45 AM · MITK

kislinsk committed rMITK3349cb419412: Merge branch 'master' into develop (authored by kislinsk).

Merge branch 'master' into develop

Jul 2 2021, 10:45 AM

kislinsk committed rMITKce85d3e8c908: 2021 Week 26 (Very Early July) (authored by kislinsk).

2021 Week 26 (Very Early July)

Jul 2 2021, 10:44 AM

kislinsk edited the content of 2021 Week 26 (Very Early July).

Jul 2 2021, 10:36 AM

kislinsk edited the content of 2021 Week 17 (Very Late April).

Jul 2 2021, 9:45 AM

kislinsk added a comment to T28602: 2021 Week 26 (Very Early July).

Pushed new branch to rMITK MITK: release/T28602-2021-Week-26.

Jul 2 2021, 9:42 AM · MITK

kislinsk triaged T28602: 2021 Week 26 (Very Early July) as Normal priority.

Jul 2 2021, 9:40 AM · MITK

kislinsk created T28602: 2021 Week 26 (Very Early July).

Jul 2 2021, 9:40 AM · MITK

kislinsk created an object: 2021 Week 26 (Very Early July).

Jul 2 2021, 9:40 AM

Jul 1 2021

kislinsk closed T28491: [Segmentation] Slice interpolation 'confirm for all slices' sometimes crahes as Resolved.

I will create another task for an optimization of the diff image operations in the do/undo stack.

Jul 1 2021, 10:10 AM · MITK (v2023.12)

kislinsk added a comment to T28491: [Segmentation] Slice interpolation 'confirm for all slices' sometimes crahes.

Deleted branch from rMITK MITK: bugfix/T28491-SpeedUp2DInterpolationsOfLargeSegmentations.

Jul 1 2021, 10:08 AM · MITK (v2023.12)

kislinsk committed rMITK014bc90c3796: Cleanup QmitkSlicesInterpolator::AcceptAllInterpolations() (authored by kislinsk).

Cleanup QmitkSlicesInterpolator::AcceptAllInterpolations()

Jul 1 2021, 10:08 AM

kislinsk committed rMITKc92776202a35: Clean up mitk::SegmentationInterpolationController::Interpolate() (authored by kislinsk).

Clean up mitk::SegmentationInterpolationController::Interpolate()

Jul 1 2021, 10:08 AM

kislinsk committed rMITK83e828cf2778: Implement simple caching of distance images (authored by kislinsk).

Implement simple caching of distance images

Jul 1 2021, 10:08 AM

kislinsk committed rMITK7814683ed6f3: Add optional parameter to allow caching of calculations (authored by kislinsk).

Add optional parameter to allow caching of calculations

Jul 1 2021, 10:08 AM

kislinsk committed rMITK72a664bc86d6: Cache some calculations for repeated 2d-interpolation (authored by kislinsk).

Cache some calculations for repeated 2d-interpolation

Jul 1 2021, 10:08 AM

kislinsk committed rMITK1c0d3e8cba5e: Derive cache size from number of H/W threads (authored by kislinsk).

Derive cache size from number of H/W threads

Jul 1 2021, 10:08 AM

kislinsk committed rMITK7f8aff5a2781: Implement optional slice extraction cache for lower and upper slices (authored by kislinsk).

Implement optional slice extraction cache for lower and upper slices

Jul 1 2021, 10:08 AM

kislinsk committed rMITK3efa9373b497: Use slice extraction cache for repeated 2d-interpolations (authored by kislinsk).

Use slice extraction cache for repeated 2d-interpolations

Jul 1 2021, 10:08 AM

kislinsk committed rMITKedebc008c3c9: Secure caches with mutexes for safe multithreading (authored by kislinsk).

Secure caches with mutexes for safe multithreading

Jul 1 2021, 10:08 AM

kislinsk committed rMITK4b229dc7cf62: Distribute slices for interpolation to multiple threads (authored by kislinsk).

Distribute slices for interpolation to multiple threads

Jul 1 2021, 10:08 AM

kislinsk committed rMITKda3a0c2eb252: Fix atomic initialization error (authored by kislinsk).

Fix atomic initialization error

Jul 1 2021, 10:08 AM

kislinsk committed rMITKe74273d042e5: Fix incomplete type error (authored by kislinsk).

Fix incomplete type error

Jul 1 2021, 10:08 AM

kislinsk committed rMITKcf9b6e2117dd: Fix unused variable (authored by kislinsk).

Fix unused variable

Jul 1 2021, 10:07 AM

kislinsk committed rMITK90a074d4c276: Fix review remarks (authored by kislinsk).

Fix review remarks

Jul 1 2021, 10:07 AM

kislinsk committed rMITKb43019736082: Speed up "Accept all slices" 2d-interpolations of segmentations (authored by kislinsk).

Speed up "Accept all slices" 2d-interpolations of segmentations

Jul 1 2021, 10:07 AM

Jun 30 2021

kislinsk added a comment to T28491: [Segmentation] Slice interpolation 'confirm for all slices' sometimes crahes.

On my 4 cores / 8 threads workstation, I get a speedup factor of nearly 10x (or 35x, if neglecting the whole do/undo stuff during measurement). This is for a 1024³ block, where only the first and last slices contain segmentations, so the whole image block need to get interpolated.

Jun 30 2021, 3:12 PM · MITK (v2023.12)

Jun 29 2021

kislinsk added a comment to T28491: [Segmentation] Slice interpolation 'confirm for all slices' sometimes crahes.

1024³ image block 2d-interpolation
==================================
Without optimizations      392.78s
Distance image cache        53.17s
-- Speedup factor ----------- 7.4x
Distance image cache +
Extract slice cache         47.22s
-- Speedup factor ----------- 8.3x

Jun 29 2021, 5:23 PM · MITK (v2023.12)

kislinsk added a comment to T28491: [Segmentation] Slice interpolation 'confirm for all slices' sometimes crahes.

Interpolation is not done in place to build a Do and Undo operation for the Undo/Redo-Stack. In the test case above this is another 30 seconds (more than half of the time with the first optimizations). While undo/redo is important, we should consider coming up with an heuristic, or option, or something similar to disable it for the interpolation of large images.

Jun 29 2021, 12:59 PM · MITK (v2023.12)

kislinsk updated subscribers of T28491: [Segmentation] Slice interpolation 'confirm for all slices' sometimes crahes.

Jun 29 2021, 12:45 PM · MITK (v2023.12)

kislinsk added a comment to T28491: [Segmentation] Slice interpolation 'confirm for all slices' sometimes crahes.

Without looking into parallelizing the interpolation yet, I was already able to algorithmically speed up the interpolation by 7-8 times.

Jun 29 2021, 12:45 PM · MITK (v2023.12)

kislinsk added a comment to T28491: [Segmentation] Slice interpolation 'confirm for all slices' sometimes crahes.

Pushed new branch to rMITK MITK: bugfix/T28491-SpeedUp2DInterpolationsOfLargeSegmentations.

Jun 29 2021, 12:42 PM · MITK (v2023.12)