• Specifications
• ImageSegmentation

TableOfContents()

Target Outcome

This specification targets at volumetric reconstruction of a tissue of interest, specifically the definition of the boundaries of the tissue,

• as an image volume, and
• as a surface representation.

Prerequisites

Include(Infrastructure/AuxiliarySoftware, "Infrastructure", 2, from="= Image Analysis & Segmentation =", to="= Geometry & Mesh Generation & Manipulation =")

Previous Protocols

For more details, see ["Specifications/ExperimentationAnatomicalImaging"].

Protocols

Input

Set(s) of MRI in NIFTI format

Segmentation using Slicer

Overview

• Download desired image volume for the desired specimen (NifTI, .nii).
• Load the desired image volume into Slicer.

• Segment the tissue of interest using the following detailed segmentation procedures can be segmented by manually painting over its boundaries or using automated or semi-automated algorithms available in the image segmentation & analysis software.

  • NOTES:
  • Save the segmented label map regularly to avoid losing tedious work through program crashes.ation in process should be saved in the relevant repository area specific to the specimen and segmentation results. The file format should be NIFTI and file naming convention should rely on tissue labeling described in ["Specifications/DataManagement"], e.g., oks001-fmb-01.nii, indicating the segmentation for femur bone of the Open Knee(s) - Specimen 1, the appended number indicates the segmentation label (to accommodate multiple segmentations of the same volume by others).
  • Segmented volume can be overlayed on other image sets to evaluate boundaries, however it only be further edited on image volumes with the same resolution as the image from which it was originally defined.
• After completion of the segmentation, the volume should be exported as an image volume (NIFTI) and a triangulated surface (without smoothing) in STL format, e.g. oks001-fmb-01.stl

Image/Data Input/Output Procedures

Load MR image data (NifTI):

• NifTI:

  • File -> Add Data -> [Select MRI file in NifTI format]

  • Open drop-down menu from toolbar for list of modules.
  • Select Editor to edit the labelmap (a volume representation of highlighted regions).
  • One can manually segment bones for every slice, or use Growth Cut Effect and Threshold Effect.
• DICOM:

  • File -> DICOM, or click DCM icon

  • Click Import
  • Select directory containing DICOM image slices (.IMA)
  • Click 'Add link'
  • CLick 'OK' after images load
  • Select desired Series in DICOM Browser window
  • Click 'Load'

Load segmented label map image data (NifTI):

• File -> Add Data

• Click 'Choose File(s) to Add'
• Select Segmented label map image (NifTI, .nii)
• Click 'Open'
• NOTE: Link 2D slice views in drop-down tack menu to ensure same label map is displayed in all views.
• NOTE: In Editor, make sure 'Merge Volume' is set to label map if applying grow-cut or other editor methods.

Save segmented label map (as NifTI):

• File -> Save, or Save icon

• Deselect all modified/selected files in the 'File Name' field
• Select/Check the desired segmented label map volume (NRRD, .nrrd)
• Set the 'File Format' to NifTI (.nii)
• Rename the label map volume as desired, see naming convention: ["Specifications/DataManagement"]
• Click 'Save'

Save triangulated surface (a.k.a. Slicer Model):

• File -> Save, or Save icon

• Deselect all modified/selected files in the 'File Name' field
• Select/Check the desired triangulated surface (VTK, .vtk)
• Set the 'File Format' to STL (.stl)
• Rename the triangulated surface as desired
• Click 'Save'

Segmentation Procedures

Manual Segmentation

Insert Katie's instructions here!

Grow-Cut Segmentation

• Editor Module
  • Applying Growth Cut Effect (region growing algorithm using cellular automata):
    • Use Draw Effect or Paint Effect to highlight anatomy of interest (foreground) using specific label (label1).
    • Use Draw Effect or Paint Effect to highlight area around anatomy of interest (background) with different label (label2).
    • Use Paint Effect to "cap" boundaries of anatomy of interest at two slices (slice before anatomy of interest appears on MRI and slice after anatomy of interest appears on MRI).

      • See figures below for example of segmentation using GrowthCut Effect with foreground (green) and background (yellow).

attachment:gc1.png attachment:gc2.png attachment:gc3.png attachment:gc4.png

• Repeat for five slices at varying depths for each view (coronal, sagittal, axial).
• Use Growth Cut Effect to complete segmentation of bone. (Apply)
• Remove background by using Change Label Effect; select label2 as input and background [0] as output. (Apply)
• Manually segmentation to clean up results of Growth Cut Effect.

• ** For more information on Growth Cut Effect refer to the 3DSlicerWiki [http://www.slicer.org/slicerWiki/index.php/Modules:GrowCutSegmentation-Documentation-3.6].

• Generating raw volume from label for anatomy of interest.
  • In Make Model Effect, choose model name, deselect smooth model and select apply.
  • Creates vtk file.
  • To save as surface representation, save vtk file as STL file.

• Tips and Shortcuts using 3DSlicer:
  • In Editor Mode:
    • 'c' to open list of labels, type # of label id and enter to switch labels quickly
    • 'z' to undo
    • 'y' to redo
    • 'd' for Draw Effect
    • 'p' for Paint Effect
    • 'e' to erase from label
    • arrow keys to move forward/background a slice
    • To change transparency of label in views or switch label to outline view.
      • Hover over thumb tack in top left of view.

      • Select '>>' button.

        • Change percent opacity by changing value near eye icon: 0.0 (transparent) to 1.0 (opaque).
        • Click on button next to eye icon to switch to outline view.

Image Processing Procedures

Label Map Smoothing

Tissue-Specific Proceduress

Registration Markers

For Registration Marker segmentation, the following technique was used:

• Use the level tracing effect to fill in the registration markers in the beginning, middle, and end of it's respective appearances within the slices, for each view.
• Next fill in any gaps with the paint tool.
• Finally, on those slices, outline the area surrounding the markers with a different color and use the grow cut effect to fill in the segmentation (as described above).

Volume labels include bone type and marker location, e.g. REG-FM, where REG stands for registration, F for femur, M for medial. In result, potential labels are REG-FM, REG-FL, REG-FA, REG-FP, REG-TM, REG-TL, REG-TA, REG-TP, REG-PS, REG-PL, REG-PM.

Bones

Input: cartilage MRI

Bone segmenatation procedure:

1. Perform Grow Cut segmentation procedure on desired bone
2. Perform Label Map Smoothing to remove Grow Cut boundary noise
3. Perform manual segmentation to more accurately define bone boundary
4. Iteratively repeat last two steps until boundary is as desired NOTE: Cortical bone will appear black in MR images, so outer edge of black bone region defines the bone surface.

Output labels: FMB (femur), TBB (tibia), FBB (fibula), PTB (patella).

Cartilage

Cartilage segmentation should be informed by bone segmentation, i.e. subtraction of bone volume to trim back surface of cartilage.

Tissue labels include FMC (femoral), TBC-M (medial tibial), TBC-L (lateral tibial), PTC (patellar).

Menisci

Menisci segmentation should be informed by cartilage segmentation, i.e. subtraction of menisci volume to match cartilage boundaries.

Tissue labels include MNS-M (medial), MNS-L (lateral).

Ligaments & Tendons

Ligament segmentation should be informed by bone segmentation, i.e. subtraction of bone volume to trim insertion areas.

Tissue labels include ACL, PCL, LCL, PTL, QAT, etc.

Output

• Volume of tissue of interest as a binary image aligned with original MRI coordinate system (raw, without filtering and smoothing)
• Surface representation of tissue of interest in STL format in MRI coordinate system (raw, without filtering and smoothing)