|
Notes:
Please reference this NITRC resource when publishing your
results that use IBSR
data: https://www.nitrc.org/projects/ibsr/
=---------------------------------------------------------------------=
Torsten Rohlfing has released these images into the public domain.
Any use of them should reference this paper:
T. Rohlfing, “Image similarity and tissue overlaps as
surrogates
for image registration accuracy: Widely used but unreliable,” IEEE
Transactions on Medical Imaging, vol. 31, no. 2, pp. 153-163, 2012.
=---------------------- original README.txt
--------------------------=
These data are
* (c) Copyright 2003-2004 The General Hospital Corporation *
* All Rights Reserved *
PLEASE REFER ALL REQUESTS FOR COPIES OF THIS DATA DIRECTLY TO THE
IBSR:
http://www.cma.mgh.harvard.edu/ibsr/
BY USING THESE DATA, YOU AGREE TO 1) ACKNOWLEDGE that the data came
from
the ibsr, and 2) PARTICIPATE in the ibsr by providing your
resulting
segmentations for comparison and by providing (at least) a
reference to
a description of your methods.
When publishing results using the data, acknowledge the source by
including
the statement, "The MR brain data sets and their manual
segmentations were
provided by the Center for Morphometric Analysis at Massachusetts
General
Hospital and are available at http://www.cma.mgh.harvard.edu/ibsr/."
The purpose of this policy is to increase collaboration. These data
have been made available for research in automating segmentation.
Specifically, you may not use and may not allow these data to be
used
for anatomical or morphological studies without explicit permission
from
the copyright holder. You must also obtain permission before
publishing
any image or segmentation data.
=---------------------------------------------------------------------=
This README is for IBSR V2.0, the 1.5mm data distribution.
For each subject we have:
T1-weighted volumetric images that have been 'positionally
normalized' into the Talairach orientation (rotation only).
Also note that these data have been processed by the CMA
'autoseg' biasfield correction routines.
Segmentation results for 'general segmentation' of the
principle gray and white matter structures of the brain
including the: 3rd Ventricle, 4th Ventricle, Brain Stem,
and Left and Right: Accumbens area , Amygdala, Amygdala Anterior,
Caudate, Cerebellum Cortex, Cerebellum Exterior, Cerebellum White
Matter, Cerebral Cortex, Cerebral Exterior, Cerebral White Matter,
Hippocampus, Inf Lat Vent, Lateral Ventricle, Pallidum, Putamen,
Thalamus Proper, VentralDC, and vessel.
Image Resolutions are as follows:
# X Y SLICE
-- ------------- ------------- -----
01 0.9375 0.9375 1.5
02 0.9375 0.9375 1.5
03 0.9375 0.9375 1.5
04 0.9375 0.9375 1.5
05 0.9375 0.9375 1.5
06 0.9375 0.9375 1.5
07 1.0 1.0 1.5
08 1.0 1.0 1.5
09 1.0 1.0 1.5
10 1.0 1.0 1.5
11 1.0 1.0 1.5
12 1.0 1.0 1.5
13 0.9375 0.9375 1.5
14 0.9375 0.9375 1.5
15 0.837053571429 0.837053571429 1.5
16 0.837053571429 0.837053571429 1.5
17 0.837053571429 0.837053571429 1.5
18 0.837053571429 0.837053571429 1.5
The images are in two formats. We include 'CMA-style'
data (ID_*.img) and Analyze (ID_ana.img) versions of the data.
CMA image files include a ASCII readable 'description file'
(ID.info) which includes, amoungst other things, the data
dimensionality, size, spatial resolutions, orientation, etc.
This information should be sufficient for image data display.
Segmentation files are also provided in the 'Native CMA'
(ID_*.otl) outline format and as a filled volume in analyze
format. The CMA .otl format is described as an appendix below.
For the filled analyze formats, the fill code represents the
various structures thta were segmented. The code-to-structure
conversion is provided in another appendix below. Also
included are "trinary" representations of the segmentations.
These are in Analyze format files where the voxel value has
been mapped from the code-to-structure codes into the basic
tissue types: backround, CSF, Gray Matter and White Matter.
The codes are converted to the numbers 0, 1, 2, and 3 as
described in the file tr_CMA_to_3 (attached below as an
appendix).
The Distribution:
There are 18 subjects, named IBSR_01 - IBSR_18. Each
subject has two gzipped tar files, one for the CMA
format and one for the Analyze format, e.g., called:
IBSR_01_CMA.tgz
IBSR_01_ANALYZE.tgz
Use the GNU version of tar to unzip and extract the tar
file using:
% tar zxvf IBSR_01_CMA.tgz
% tar zxvf IBSR_01_ANALYZE.tgz
This will create the IBSR_01 directory with the following
sub-directory structure:
$ ls IBSR_01/*/*
IBSR_01/images/analyze:
IBSR_01_ana.hdr IBSR_01_ana.img IBSR_01_ana.mat
IBSR_01/images/CMA:
IBSR_01_100.img IBSR_01_14.img IBSR_01_44.img IBSR_01_74.img
IBSR_01_101.img IBSR_01_15.img IBSR_01_45.img IBSR_01_75.img
IBSR_01_102.img IBSR_01_16.img IBSR_01_46.img IBSR_01_76.img
...
IBSR_01_126.img IBSR_01_3.img IBSR_01_6.img IBSR_01_9.img
IBSR_01_127.img IBSR_01_40.img IBSR_01_70.img IBSR_01.info
IBSR_01_128.img IBSR_01_41.img IBSR_01_71.img
IBSR_01_12.img IBSR_01_42.img IBSR_01_72.img
IBSR_01_13.img IBSR_01_43.img IBSR_01_73.img
BSR_01/segmentation/analyze:
IBSR_01_seg_ana.hdr IBSR_01_seg_ana.mat IBSR_01_segTRI_ana.img
IBSR_01_seg_ana.img IBSR_01_segTRI_ana.hdr IBSR_01_segTRI_ana.mat
IBSR_01/segmentation/CMA:
IBSR_01_seg_100.otl IBSR_01_seg_29.otl IBSR_01_seg_65.otl
IBSR_01_seg_101.otl IBSR_01_seg_2.otl IBSR_01_seg_66.otl
IBSR_01_seg_102.otl IBSR_01_seg_30.otl IBSR_01_seg_67.otl
...
IBSR_01_seg_25.otl IBSR_01_seg_61.otl IBSR_01_seg_98.otl
IBSR_01_seg_26.otl IBSR_01_seg_62.otl IBSR_01_seg_99.otl
IBSR_01_seg_27.otl IBSR_01_seg_63.otl IBSR_01_seg_9.otl
IBSR_01_seg_28.otl IBSR_01_seg_64.otl
The largest compressed tar file is almost 5MB, all compressed tar
files
require just over 62MB, each uncompressed directory containing
both Analyze and CMA data is around 136MB and all of the
uncompressed data requires 680MB.
WARNING: Dan Withey reports that there seems to be an artifact
introduced by uncompressing the ".otl" files with WinZip
Some versions of WinZip (eg. ver 9.0) will open the CMA
tar archives. For proper recovery of the mixed ascii-binary
.otl files The option "TAR file smart CR/LF conversion" under
Options-Configuration-Miscellaneous should be disabled.
# Sex age
--- ----------
1 M 37
2 M 41
3 F JUV
4 M JUV
5 M 41
6 M 46
7 F 70
8 M 60
9 M 41
10 F 35
11 F 59
12 M 71
13 M JUV
14 M JUV
15 M 8
16 M 7
17 M 8
18 M 13
GDF File Format
----------------------------------------------------------------------
GDF stands for Generic Data Format.
GDF files have a main header that defines the defaults for the rest
of
the file. The rest of the file is made up of "blocks". Each block
has
a header and data. The bulk of data should be stored in the data
portion of the blocks. The header must contain information
describing
the size and type of the bulk data, but can contain other
associated
descriptors (e.g a label) for that block. The programmer will
usually
design the data structure such that information pertinent to the
entire
series of data blocks is assigned to the main header and is absent
in
the block headers. Information retrieval routines query the main
header
after determining field is absent in the block header, so a
specified
block header value overrides the main header value. Headers are
always
ascii, both in memory and on file. The data blocks are always
non-ascii
in memory, but can be represented either as ascii or non-ascii on
file.
Here is an example (the gdf file is on the left and comments are on
the
right):
GDF FILE VERSION3 | Main header: version ID
START MAIN HEADER | Main header:
TYPE short | Main header: field 0
COL_NUM 2 | Main header: field 1
ONUM 5 | Main header: ...etc...
END MAIN HEADER | Main header:
START GDF HEADER | BLock 0 header:
ROW_NUM 248 | Block 0 header: field 0
SEED 309 301 | Block 0 header: field 1
LABEL Tumor Boundary | Block 0 header: ...etc...
END GDF HEADER | Block 0 header:
START POINTS | Block 0 data:
| Block 0 data: bulk data here
END POINTS | Block 0 data:
START GDF HEADER | BLock 1 header:
ROW_NUM 248 | Block 1 header: field 0
SEED 309 301 | Block 1 header: field 1
LABEL Tumor Boundary | Block 1 header: ...etc...
END GDF HEADER | Block 1 header:
START POINTS | Block 1 data:
| Block 0 data: bulk data here
END POINTS | Block 1 data:
START GDF HEADER | ...etc...
Header fields are stored as text and consist of:
All gdf files will have the following fields:
TYPE
COL_NUM
ROW_NUM
Bulk data are stored in a COL_NUM by ROW_NUM array and
have a that can be any C data type:
char
short
double
float
long
unsigned char
unsigned short
table /* gdf specific type for using mulitple C data types */
NOTE: All numbers in gdf files are assumed to be big endian since
gdf
was originally implemented on a Sun. This means the most
significant
byte of multiple-byte words are stored first (lower addresses).
Fill code to Structure List for Segmentation
----------------------------------------------------------------------
# tkmeditColorsCMA from the FreeSurfer distribution
#
# Format:
# Fill_Value Anatomic_Label RGB Color values Alpha_channel
#
#
0 Unknown 0 0 0 0
1 Left-Cerebral-Exterior 205 62 78 0
2 Left-Cerebral-White-Matter 225 225 225 0
3 Left-Cerebral-Cortex 205 62 78 0
4 Left-Lateral-Ventricle 120 18 134 0
5 Left-Inf-Lat-Vent 196 58 250 0
6 Left-Cerebellum-Exterior 0 148 0 0
7 Left-Cerebellum-White-Matter 220 248 164 0
8 Left-Cerebellum-Cortex 230 148 34 0
9 Left-Thalamus 0 118 14 0
10 Left-Thalamus-Proper 0 118 14 0
11 Left-Caudate 122 186 220 0
12 Left-Putamen 236 13 176 0
13 Left-Pallidum 12 48 255 0
14 3rd-Ventricle 204 182 142 0
15 4th-Ventricle 42 204 164 0
16 Brain-Stem 119 159 176 0
17 Left-Hippocampus 220 216 20 0
18 Left-Amygdala 103 255 255 0
19 Left-Insula 80 196 98 0
20 Left-Operculum 60 58 210 0
21 Line-1 60 58 210 0
22 Line-2 60 58 210 0
23 Line-3 60 58 210 0
24 CSF 60 60 60 0
25 Left-Lesion 255 165 0 0
26 Left-Accumbens-area 255 165 0 0
27 Left-Substancia-Nigra 0 255 127 0
28 Left-VentralDC 165 42 42 0
29 Left-undetermined 135 206 235 0
30 Left-vessel 160 32 240 0
31 Left-choroid-plexus 0 255 255 0
32 Left-F3orb 100 50 100 0
33 Left-lOg 135 50 74 0
34 Left-aOg 122 135 50 0
35 Left-mOg 51 50 135 0
36 Left-pOg 74 155 60 0
37 Left-Stellate 120 62 43 0
38 Left-Porg 74 155 60 0
39 Left-Aorg 122 135 50 0
40 Right-Cerebral-Exterior 205 62 78 0
41 Right-Cerebral-White-Matter 0 225 0 0
42 Right-Cerebral-Cortex 205 62 78 0
43 Right-Lateral-Ventricle 120 18 134 0
44 Right-Inf-Lat-Vent 196 58 250 0
45 Right-Cerebellum-Exterior 0 148 0 0
46 Right-Cerebellum-White-Matter 220 248 164 0
47 Right-Cerebellum-Cortex 230 148 34 0
48 Right-Thalamus 0 118 14 0
49 Right-Thalamus-Proper 0 118 14 0
50 Right-Caudate 122 186 220 0
51 Right-Putamen 236 13 176 0
52 Right-Pallidum 255 48 255 0
53 Right-Hippocampus 220 216 20 0
54 Right-Amygdala 103 255 255 0
55 Right-Insula 80 196 98 0
56 Right-Operculum 60 58 210 0
57 Right-Lesion 255 165 0 0
58 Right-Accumbens-area 255 165 0 0
59 Right-Substancia-Nigra 0 255 127 0
60 Right-VentralDC 165 42 42 0
61 Right-undetermined 135 206 235 0
62 Right-vessel 160 32 240 0
63 Right-choroid-plexus 0 255 255 0
64 Right-F3orb 100 50 100 0
65 Right-lOg 135 50 74 0
66 Right-aOg 122 135 50 0
67 Right-mOg 51 50 135 0
68 Right-pOg 74 155 60 0
69 Right-Stellate 120 62 43 0
70 Right-Porg 74 155 60 0
71 Right-Aorg 122 135 50 0
72 5th-Ventricle 120 190 150 0
73 Right_Thin_Cerebral_White_Matter 240 240 240 0
74 Left_Thin_Cerebral_White_Matter 240 240 240 0
80 Bone 100 100 100 0
81 Fat 255 255 255 0
82 Bright-Unknown 100 240 240 0
83 Dark_Unknown 20 100 100 0
84 Corpus_Callosum 50 255 150 0
Fill code to Structure List for Parcellations
----------------------------------------------------------------------
# tkmeditLatParcColorsCMA
#
# Format:
# Fill_Value Anatomic_Label RGB Color values Alpha_channel
#
0 Unknown 0 0 0 0
1 Dummy 0 0 0 0
2 Left-FP 0 238 0 0
3 Left-INS 160 82 45 0
4 Left-F1 255 218 185 0
5 Left-F2 0 206 209 0
6 Left-F3t 127 255 212 0
7 Left-F3o 178 34 34 0
8 Left-PRG 238 0 0 0
9 Left-TP 34 139 34 0
10 Left-T1a 208 32 144 0
11 Left-T1p 173 255 47 0
12 Left-T2a 240 230 140 0
13 Left-T2p 173 216 230 0
14 Left-TO2 238 238 0 0
15 Left-T3a 50 205 50 0
16 Left-T3p 255 0 255 0
17 Left-TO3 176 48 96 0
18 Left-POG 0 255 127 0
19 Left-SPL 245 222 179 0
20 Left-SG 255 165 5 0
21 Left-SGa 255 165 0 0
22 Left-SGp 255 69 0 0
23 Left-AG 205 91 69 0
24 Left-OLs 255 192 203 0
25 Left-OLi 152 251 152 0
26 Left-CALC 100 149 237 0
27 Left-FMC 160 32 240 0
28 Left-SMC 238 130 238 0
29 Left-SC 238 201 0 0
30 Left-PAC 218 112 214 0
31 Left-CGa 255 62 150 0
32 Left-CGp 0 0 255 0
33 Left-PCN 39 64 139 0
34 Left-CN 250 128 114 0
35 Left-FOC 255 110 180 0
36 Left-PHa 255 99 71 0
37 Left-PHp 255 255 0 0
38 Left-LG 0 100 0 0
39 Left-TFa 205 92 92 0
40 Left-TFp 165 42 42 0
41 Left-TOF 153 50 204 0
42 Left-OF 0 255 255 0
43 Left-FO 221 160 221 0
44 Left-CO 135 206 235 0
45 Left-CaO 152 200 214 0
46 Left-CpO 153 29 242 0
47 Left-PO 210 180 140 0
48 Left-PP 255 215 0 0
49 Left-H1 0 0 128 0
50 Left-PT 46 139 87 0
51 Left-SCLC 102 205 170 0
52 Left-OP 0 255 0 0
53 Left-rm 0 0 0 0
54 Left-Hippocampus 220 216 20 0
55 Left-Amygdala 60 58 210 0
56 Left-F3orb 100 50 100 0
57 Left-lOg 135 50 74 0
58 Left-aOg 122 135 50 0
59 Left-mOg 51 50 135 0
60 Left-pOg 74 155 60 0
61 Right-FP 0 238 0 0
62 Right-INS 160 82 45 0
63 Right-F1 255 218 185 0
64 Right-F2 0 206 209 0
65 Right-F3t 127 255 212 0
66 Right-F3o 178 34 34 0
67 Right-PRG 238 0 0 0
68 Right-TP 34 139 34 0
69 Right-T1a 208 32 144 0
70 Right-T1p 173 255 47 0
71 Right-T2a 240 230 140 0
72 Right-T2p 173 216 230 0
73 Right-TO2 238 238 0 0
74 Right-T3a 50 205 50 0
75 Right-T3p 255 0 255 0
76 Right-TO3 176 48 96 0
77 Right-POG 0 255 127 0
78 Right-SPL 245 222 179 0
79 Right-SG 255 165 5 0
80 Right-SGa 255 165 0 0
81 Right-SGp 255 69 0 0
82 Right-AG 205 91 69 0
83 Right-OLs 255 192 203 0
84 Right-OLi 152 251 152 0
85 Right-CALC 100 149 237 0
86 Right-FMC 160 32 240 0
87 Right-SMC 238 130 238 0
88 Right-SC 238 201 0 0
89 Right-PAC 218 112 214 0
90 Right-CGa 255 62 150 0
91 Right-CGp 0 0 255 0
92 Right-PCN 39 64 139 0
93 Right-CN 250 128 114 0
94 Right-FOC 255 110 180 0
95 Right-PHa 255 99 71 0
96 Right-PHp 255 255 0 0
97 Right-LG 0 100 0 0
98 Right-TFa 205 92 92 0
99 Right-TFp 165 42 42 0
100 Right-TOF 153 50 204 0
101 Right-OF 0 255 255 0
102 Right-FO 221 160 221 0
103 Right-CO 135 206 235 0
104 Right-CaO 152 200 214 0
105 Right-CpO 153 29 242 0
106 Right-PO 210 180 140 0
107 Right-PP 255 215 0 0
108 Right-H1 0 0 128 0
109 Right-PT 46 139 87 0
110 Right-SCLC 102 205 170 0
111 Right-OP 0 255 0 0
112 Right-rm 0 0 0 0
113 Right-Hippocampus 220 216 20 0
114 Right-Amygdala 60 58 210 0
115 Right-F3orb 100 50 100 0
116 Right-lOg 135 50 74 0
117 Right-aOg 122 135 50 0
118 Right-mOg 51 50 135 0
119 Right-pOg 74 155 60 0
120 Medial 1 1 1 0
tr_CMA_to_3
----------------------------------------------------------------------
0 0
1 0
2 3
3 2
4 1
5 1
6 0
7 3
8 2
9 2
10 2
11 2
12 2
13 2
14 1
15 1
16 3
17 2
18 2
19 2
20 2
21 0
22 0
23 0
24 1
25 0
26 2
27 2
28 2
29 0
30 0
31 0
32 0
33 0
34 0
35 0
36 0
37 0
38 0
39 0
40 0
41 3
42 2
43 1
44 1
45 0
46 3
47 2
48 2
49 2
50 2
51 2
52 2
53 2
54 2
55 2
56 2
57 0
58 2
59 2
60 2
61 0
62 0
63 0
64 0
65 0
66 0
67 0
68 0
69 0
70 0
71 0
72 0
73 0
74 0
80 0
81 0
82 0
83 0
84 3
Changes:
|
