MITK Segmentation Redesign

Segmentation Redesign

Since the last release from 2013.06 the whole segmentation group has been working on the extensive redesign of our segmentation plugin. With the MITK 2013.09 release we are proud to announce that this project is completed! Here is a list of the major changes and new features:

General

• We cleaned up the segmentation view and removed unused GUI elements
• New button icons
• The toolmanager can now be accessed using CppMicroServices
• Switch to mitkContourModel:
  • All 2D tools are now using the new mitkContourModel
  • There is also a new data structure called mitkContourModelSet which allow the storage of a set of ContourModels
  • The old mitkContour and mitkContourSet are declared as deprecated and will be removed during the next year
• 3D interpolation can now also be used with planar figures
• New !SegmentationUI module which contains all the segmentation widgets
• Many new tools and utilities

Segmentation View

• It contains now two tabs which allows to switch between 2D and 3D tools
• We extended the 2D tool set. There is now also a 2D Fast Marching Tool
• The 3D tool tab contain many tools which were formerly located in the data manager context menu. There are also several new tools like
  • Upper/Lower Threshold
  • Fast Marching 3D
  • Watershed Segmentation
• We also improved the GUI of all of the 3D tools, especially
  • 3D Regiongrowing
  • Otsu Segmentation
• Improved the interpolation widget
  • By default interpolation is disabled at the moment

Segmentation Utilities

Often a segmentation needs some post processing. Lets say you just performed a threshold segmentation and the result contains some "holes". In this case you would like to fill these holes automatically. Or you want to combine to different segmentation into one single segmentation. For these cases MITK now comes with a extensive set of segmentation utilities in a separate view:

• Boolean operations:
  • Calculate the difference between two segmentations
  • Calculate the intersection of two segmentations
  • Calculate the union of the segmentations
• Image masking:
  • You have finished a segmentation and what now? E.g. you can "mask" the patient image with you segmentation or a surface and then proceed with e.g. vessel segmentation/visualization
• Morphological operations:
  • Postprocess your segmentation by
    • Dilatation/Erosion
    • Opening/Closing
    • Fill Holes