CIVET stage
9. nlfit
10. mask_classify
Kwon, hun ki
Stages
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
nuc_t1_native
skull_masking_native
stx_register
stx_tal_to_7
stx_tal_to_6
tal_t1
nuc_inorm_t1
skull_removal
nlfit
mask_classify
pve_curvature
pve
reclassify
segment
cls_volumes
cortical_masking
segment_volumes
surface_classify
artefact
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
create_wm_hemispheres
segment_mask
expand_from_white_left
expand_from_white_right
slide_left_hemi_obj_back
flip_right_hemi_obj_back
slide_right_hemi_obj_back
calibrate_left_white
calibrate_right_white
laplace_field
gray_surface_left
gray_surface_right
mid_surface_left
mid_surface_right
surface_fit_error
verify_image_nlfit
gyrification_index_left
verify_brain_mask
classify_qc
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
dataterm_left_surface
brain_mask_qc
gyrification_index_right
dataterm_right_surface
surface_registration_left
surface_registration_right
mean_curvature_20mm_left
mean_curvature_20mm_right
thickness_tlink_20mm_right
thickness_tlink_20mm_left
resample_left_mean_curvature
resample_right_mean_curvature
resample_right_thickness
resample_left_thickness
lobe_area_right
lobe_area_left
verify_clasp
verify_image
9.nlfit
•
•
•
•
•
Label: "creation of nonlinear transform"
실행명령:
 /progs/best1stepnlreg.pl(s) [options] source.mnc target.mnc output.xfm
[output.mnc]
 mincresample
 inormalize
 minccalc
 mincblur
Input : mni_icbm_00340_t1_final.mnc
Output : mni_icbm_00340_t1_nlfit_lt.xfm
best1stepnlreg.pl does hierachial non-linear fitting between two files
you will have to edit the script itself to modify the fitting levels
themselves
9.nlfit
nlfit.log
9.nlfit
Results
T1_final
icbm152
T1_to_icbm
(nonlinear)
T1_to_icbm
on T1_final
10. mask_classify
•
Label: “tissue classification”
•
실행명령:

classify_clean
 transform_tags
 classify
 Clean_tag
•
Input : mni_icbm_00340_t1_final.mnc
mni_icbm_00340_skull_mask.mnc
•
Output : mni_icbm_00340_clean.mnc
•
의존성 : nlfit
Steps involved:
1. call classify -min_dist to compute initial classification
2. correct standard tag file with min_dist classification
3. call classify -ann to compute final classification
10. mask_classify
classify_clean [options] <in.mnc> [<in.mnc> ...] <classified.mnc>
classify_clean -clobber -clean_tags -mask_classified -mask_tag
-mask mni_icbm_00340_skull_mask.mnc -tag_transform mni_icbm_00340_t1_nlfit_lt.xfm
mni_icbm_00340_t1_final.mnc mni_icbm_00340_clean.mnc
<option>
–
–
–
–
–
–
-clean_tags : clean tag file using mindist pre-classification [default is -noclean_tags]
-mask_classified apply mask to intermediate and final classified volumes [default: nomask_classified]
-mask <mask.mnc> : specify a mask volume
-maskbinvalue : value of mask foreground [default: assume a binary mask]
-mask_tag : apply mask to foreground tags prior to classification(s) [default: -nomask_tag]
-tag_transform : non-linear transformation to map the tags from stereotaxic space to
subject
10. mask_classify
Results
Sub-routine
 transform_tags
Usage: transform_tags input.tag input.xfm [output.tag] [invert]
transform_tags ntags_1000_prob_90_nobg.tag
mni_icbm_00340_nlfit_It.xfm output.tag invert
– Transforms the input tags by the input transform. If a fourth argument is
present, then the inverse of the transform is used. The transformed tags are
written to output.tag if specified, otherwise input.tag is overwritten.
Sub-routine
input.tag & output.tag
Result
Sub-routine
 clean_tag
cleantag -oldtag /out_tag.tag -newtag /masked_standard.tag
-mask /mni_icbm_00340_skull_mask.mnc -maskbinvalue 1
< Option >
– oldtag : Specify the tag file to be cleaned
– newtag : Specify the file name of the clean set of tag points
– mask : Specify a mask to apply to the tag points
– maskbinvalue : Value of mask foreground Default value: 1
Sub-routine
masked_standard.tag
Result
Sub-routine
 classify
Usage: classify <options> <infile1> [infile2] ... <outfile>
classify -verbose -clobber -mask mni_icbm_00340_skull_mask.mnc
-user_mask_value 0.5
-min -tagfile masked_standard.tag
-fuzzy all mni_icbm_00340_t1_final.mnc
mni_icbm_00340_t1_final_fuzzy.mnc
< Option >
– min : Use the 'Minimum Distance' classifier.
– user_mask_value: Specify the mask value. (If the mask is a
classified volume) Default value: 1
– tagfile: Input training points as tag file (.tag)
– Fuzzy : Specify a string ex. '011...' to indicate classes for fuzzy
classification.
Sub-routine
Results
Sub-routine
Results
Sub-routine
 clean_tag
Usage: cleantag [options] <fuzzy_class.mnc> <class_id> [<fuzzy_class.mnc> <class_id> ...]
cleantag -oldtag /masked_standard.tag
–newtag mni_icbm_00340_t1_final_fuzzy_cleaned.tag -mode 110
-threshold 0.7 -difference 0.3
-comment '-mode 110 -threshold 0.7 -difference 0.3'
/mni_icbm_00340_t1_final_fuzzy_1.mnc 1
/mni_icbm_00340_t1_final_fuzzy_2.mnc 2
/mni_icbm_00340_t1_final_fuzzy_3.mnc 3
< Option >
–
oldtag : Specify the tag file to be cleaned
–
newtag : Specify the file name of the clean set of tag points
–
Mode : Specify a quoted string 'xyz' to denote tag rejection mode.
–
Each of x,y,z being 0 or 1, effects are additive.
–
–
–
–
–
x = 1, if tag class label <> voxel class (highest fuzzy voxel class),
y = 1, if fuzzy voxel class < threshold,
z = 1, if diff. between 2 highest fuzzy voxel classes < diff. threshold,
Threshold : Set the fuzzy threshold rejection criterion
Difference : Set the difference threshold rejection criterion
Sub-routine
Results
Sub-routine
cleantag -oldtag mni_icbm_00340_t1_final_fuzzy_cleaned.tag -newtag
/masked_custom.tag -mask mni_icbm_00340_skull_mask.mnc maskbinvalue 1
Sub-routine
Results
Sub-routine
classify -verbose -clobber -mask mni_icbm_00340_skull_mask.mnc
-user_mask_value 0.5 -ann
–tagfile masked_custom.tag %%%mni_icbm_00340_t1_final_fuzzy_custom.tag
mni_icbm_00340_t1_final.mnc mni_icbm_00340_cls_clean.mnc
< Option >
-ann: Use the 'Artificial Neural Network' classifier.
Sub-routine
Results