Difference between revisions of "Smoothing multi-label brain segmentations"

Revision as of 22:31, 27 January 2010

Home < Smoothing multi-label brain segmentations

  WM label 3
  CSF label 1
  GM(without the following specified sub-cortical structures) label 2,
  GM sub-cortical structures

Segmentation using one of the approaches below
Up to Affine registration of the subject to bias-corrected, skull-stripped Baseline scan
Map the Baseline ICC mask to the State 1 space
Skull stripping State 1 scan
Bias-correction of State 1 scan
Repeat up to affine registration of skull-stripped, bias-corrected State 1 scan to Baseline

@@ Line 1: / Line 1: @@
 == smoothing workflow ==
-# Deformable registration of the atlas to the subject to be segmented
+# input: a multi-label map to be smoothed. With structures:
-# Segmentation using one of the approaches below
-For State 1 scans, the workflow is the following:
+   WM label 3
+   CSF label 1
+   GM(without the following specified sub-cortical structures) label 2,
+   GM sub-cortical structures
+# Segmentation using one of the approaches below
 # Up to Affine registration of the subject to bias-corrected, skull-stripped Baseline scan
 # Map the Baseline ICC mask to the State 1 space
@@ Line 11: / Line 15: @@
 # Bias-correction of State 1 scan
 # Repeat up to affine registration of skull-stripped, bias-corrected State 1 scan to Baseline
-== Considered segmentation approaches ==
-Choice of atlas:
-* Vervet atlas
-* Rhesus atlas
-Choice of segmentation approach:
-* fluid deformable registration of a specific atlas to the subject (UNC approach)
-* EM Segmenter (Kilian Pohl)
-* itkEMS/ABC (Marcel Prastawa)