Initial atlas construction workflow - Revision history

Fedorov: moved Current atlas construction workflow to Initial atlas construction workflow: The atlas construction workflow has been revised. New workflow has been adopted.

2009-12-03T17:51:01Z

moved Current atlas construction workflow to Initial atlas construction workflow: The atlas construction workflow has been revised. New workflow has been adopted.

Fedorov: /* Workflow / Tools */

2009-10-06T01:15:04Z

Workflow / Tools

Ginger at 20:59, 5 October 2009

2009-10-05T20:59:23Z

Ginger at 20:44, 5 October 2009

2009-10-05T20:44:42Z

Ginger at 20:35, 5 October 2009

2009-10-05T20:35:03Z

Ginger at 20:34, 5 October 2009

2009-10-05T20:34:37Z

Ginger at 20:24, 5 October 2009

2009-10-05T20:24:25Z

Ginger at 19:16, 5 October 2009

2009-10-05T19:16:52Z

Ginger at 19:15, 5 October 2009

2009-10-05T19:15:39Z

Ginger: Created page with '== Objective == Develop an atlas construction workflow for vervet segmentation, which will maximally (completely) consist of the NA-MIC tools. Learn from the previous experience…'

2009-10-05T19:12:59Z

Created page with '== Objective == Develop an atlas construction workflow for vervet segmentation, which will maximally (completely) consist of the NA-MIC tools. Learn from the previous experience…'

New page

== Objective ==

Develop an atlas construction workflow for vervet segmentation, which will maximally (completely) consist of the NA-MIC tools. Learn from the previous experience and try to mimic as closely as possible the workflow developed by Styner et al. [1]

== Data ==

10 vervet subjects (2 subjects possibly have unacceptable image quality), T1 sequence

== Workflow / Tools ==

{| class="wikitable" style="text-align:center; width:800px; height:200px" border="1"
|-
! Workflow step !! Description !! Tools
|-
! 1. Preparation
|
1.1. Select randomly a template image subject

1.2. Semi-automatically identify the ICC in the template subject.

1.3. 6+9+12 DOF registration of each of the subjects to the template.

1.4. Skull-strip each subject using dilated ICC of the template.

1.5. Perform intensity calibration to the template.

1.6. Use affine-registered, skull-stripped images to form initial affine average image.
||
1.1. n/a

1.2. Slicer Editor module, SkullStripper Slicer module

1.3. Slicer BRAINSFit extension

1.4. Slicer Mask, Editor modules (verify the mask encloses actual ICC!)

1.5. Slicer Histogram Matching module

1.6. ''Tool to be developed'': compute average image from the set of input images
|-
! 2. Atlas construction
|

2.1. Use the affine registered image from Step 1.6 as the template to repeat Steps 1.1-1.6.

2.2. Perform non-rigid alignment of the subjects affinely registered to the template.

2.3. Compute the atlas as the average.

||
2.1 see above

2.2. Slicer BRAINSFit extension (BSpline), possibly Groupwise BSpline registration by Balci et al. [2]

2.3. Tool from step 1.6.

|-
! 3. Probabilistic atlas construction
|

3.1. Segment WM/GM/CSF from the averaged template, manually edit to ensure accuracy

3.2. Slightly smooth the segmentations Styner et al. recommend kernel of 0.4 mm variance.

3.3. Back-propagate WM/GM/CSF segmentations to the training subjects using the deformation field from Step 2.2.

3.4. Construct the probabilistic atlas by averaging of the tissue maps, normalize locally to 1, construct rejection class by inverting the sum of WM+GM+CSF.

|

3.1. Slicer editor module (thresholding, manual editing)

3.2. Gaussian smoothing module (check availability)

3.3. Slicer applyDeformation module (command line accessible only!)

3.4. Tool from Step 1.6, tool to be developed, Slicer Editor module

|}

== Standing questions ==

* when to perform bias correction? After Step 1.4?
* deep brain structures -- similar approach for identification in subjects as for GM/WM/CSF in Step 3.1?
* Styner et al. used T2 to segment CSF. We only have T1

== References ==

# M. Styner, R. Knickmeyer, S. Joshi, C. Coe, S. J. Short, and J. Gilmore. Automatic brain segmentation in rhesus monkeys. Proc SPIE Medical Imaging Conference, Proc SPIE Vol 6512 Medical Imaging 2007, pp 65122L-1 - 65122L-8 [http://www.google.com/url?sa=t&source=web&ct=res&cd=1&url=http%3A%2F%2Fwww.cs.unc.edu%2F~styner%2Fpublic%2FMyPapers%2Fpapers%2FMedical%2520Imaging%25202007%2520Image%2520Processing%25202007%2520Styner-1.pdf&ei=DubDSt3AAciwlAfJp5TwBA&usg=AFQjCNGojUvBkh6TwbS0vJ9X7OEFLDFxqQ&sig2=wHgXTrdAXzUC5DuL_C7CCQ pdf]
# Balci S.K., Golland P., Wells W.M. Non-rigid Groupwise Registration using B-Spline Deformation Model. Insight Journal - 2007 MICCAI Open Science Workshop. [http://hdl.handle.net/1926/568 link]
# Previous descriptions of the atlas construction workflow: [http://wiki.na-mic.org/Wiki/index.php/Atlas-to-subject_registration_with_FSL_and_Diffeomorphic_Demons,_for_EM_segmentation. Summary by Ginger Li], [http://www.bsl.ece.vt.edu/data/vervet_atlas/vervet.php Description off BSL atlas page]

@@ Line 60: / Line 60: @@
 .2. Slightly smooth the segmentations Styner et al. recommend kernel of 0.4 mm variance.
+|
@@ Line 65: / Line 67: @@
 .1. Kmeans + manual
-.2. Gaussian smoothing module in Slicer, 0.6 and 0.65 kernel size used for Valentino and Tommy, respectively.
+.2. Gaussian smoothing module in Slicer, 0.6 and 0.65 kernel size used for Valentino and Tommy, respectively. '''AF: I do not understand what is the relation of Valentino and Tommy to this, if the operation was performed on hard segmentation from the average template'''
 *We did not perform back propagation as in Styner07 due to the large structural differences in the dataset.
 |}
 == References ==

@@ Line 32: / Line 32: @@
 .3. Slicer Histogram Matching module
-.4. FSL/Flirt
+.4. FSL-Flirt
 .5. Slicer Mask, Editor modules (verify the mask encloses actual ICC!)

@@ Line 40: / Line 40: @@
+|
-.1. Use the skullstriped image from Step 1.2 as the template to repeat Steps 1.3-1.4.
+.1. Use the skullstriped image from Step 1.2 as the template to repeat Step 1.4.
 .2. Perform non-rigid alignment of the subjects affinely registered to the template.
@@ Line 51: / Line 51: @@
 .2. Diffeomorphic demons
-.3. Some tool in Slicer under develop
+.3. Some tool in Slicer under development
 |-
@@ Line 67: / Line 67: @@
 .2. Gaussian smoothing module in Slicer, 0.6 and 0.65 kernel size used for Valentino and Tommy, respectively.
-*We did not perform back propagation due to the large structural differences in the dataset.
+*We did not perform back propagation as in Styner07 due to the large structural differences in the dataset.
 |}
 == References ==

@@ Line 67: / Line 67: @@
 .2. Gaussian smoothing module in Slicer, 0.6 and 0.65 kernel size used for Valentino and Tommy, respectively.
-We did not perform back propagation due to the large structural differences in the dataset.
+*We did not perform back propagation due to the large structural differences in the dataset.
 |}

@@ Line 19: / Line 19: @@
 .2. Manually identify the ICC in the template subject.
-.3. Perform intensity calibration to the template.
+.3. Perform intensity calibration for all other subjects to the template.
 .4. 12 DOF registration of each of the subjects to the template.
@@ Line 40: / Line 40: @@
+|
-.1. Use the affine registered image from Step 1.6 as the template to repeat Steps 1.1-1.6.
+.1. Use the skullstriped image from Step 1.2 as the template to repeat Steps 1.3-1.4.
 .2. Perform non-rigid alignment of the subjects affinely registered to the template.
@@ Line 49: / Line 49: @@
 .1 see above
-.2. Slicer BRAINSFit extension (BSpline), possibly Groupwise BSpline registration by Balci et al. [2]
+.2. Diffeomorphic demons
-.3. Tool from step 1.6.
+.3. Some tool in Slicer under develop
 |-
@@ Line 60: / Line 60: @@
 .2. Slightly smooth the segmentations Styner et al. recommend kernel of 0.4 mm variance.
+|
-.1. Slicer editor module (thresholding, manual editing)
+.1. Kmeans + manual
-.3. Slicer applyDeformation module (command line accessible only!)
+.2. Gaussian smoothing module in Slicer, 0.6 and 0.65 kernel size used for Valentino and Tommy, respectively.
-.4. Tool from Step 1.6, tool to be developed, Slicer Editor module
+We did not perform back propagation due to the large structural differences in the dataset.
 |}

← Older revision	Revision as of 17:51, 3 December 2009
(No difference)

@@ Line 17: / Line 17: @@
 .1. Select the subject that is best oriented as the template image
-.2. Semi-automatically identify the ICC in the template subject.
+.2. Manually identify the ICC in the template subject.
-.3. 6+9+12 DOF registration of each of the subjects to the template.
+.3. Perform intensity calibration to the template.
-.4. Skull-strip each subject using dilated ICC of the template.
+.4. 12 DOF registration of each of the subjects to the template.
-.5. Perform intensity calibration to the template.
+.5. Skull-strip each subject using dilated ICC of the template.
 ||
 .1. n/a
-.2. Slicer Editor module, SkullStripper Slicer module
+.2. manual in SNAP.
-.3. Slicer BRAINSFit extension
+.3. Slicer Histogram Matching module
-.4. Slicer Mask, Editor modules (verify the mask encloses actual ICC!)
+.4. FSL/Flirt
-.5. Slicer Histogram Matching module
+.5. Slicer Mask, Editor modules (verify the mask encloses actual ICC!)
 |-
 ! 2. Atlas construction

@@ Line 1: / Line 1: @@
-== Objective ==
+== Content ==
-Develop an atlas construction workflow for vervet segmentation, which will maximally (completely) consist of the NA-MIC tools. Learn from the previous experience and try to mimic as closely as possible the workflow developed by Styner et al. [1]
+This page describes the current working pipeline of atlas construction for vervet data.
 == Data ==

@@ Line 15: / Line 15: @@
 ! 1. Preparation
+|
-.1. Select randomly a template image subject
+.1. Select the subject that is best oriented as the template image
 .2. Semi-automatically identify the ICC in the template subject.