Chris Oswald
commited on
Commit
·
f2ef76c
1
Parent(s):
81b3fb4
added original subset indicator
Browse files
SPIDER.py
CHANGED
@@ -73,8 +73,37 @@ _CITATION = """\
|
|
73 |
# TODO: Add description of the dataset here
|
74 |
# You can copy an official description
|
75 |
_DESCRIPTION = """\
|
76 |
-
This
|
77 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
78 |
|
79 |
_HOMEPAGE = "https://zenodo.org/records/10159290"
|
80 |
|
@@ -106,7 +135,7 @@ class CustomBuilderConfig(datasets.BuilderConfig):
|
|
106 |
|
107 |
|
108 |
class SPIDER(datasets.GeneratorBasedBuilder):
|
109 |
-
"""TODO: Short description of my dataset."""
|
110 |
|
111 |
# Class attributes
|
112 |
DEFAULT_WRITER_BATCH_SIZE = 16 # PyArrow default is too large for image data
|
@@ -174,6 +203,7 @@ class SPIDER(datasets.GeneratorBasedBuilder):
|
|
174 |
"TransmitCoilName": datasets.Value(dtype="string"),
|
175 |
"WindowCenter": datasets.Value(dtype="string"),
|
176 |
"WindowWidth": datasets.Value(dtype="string"),
|
|
|
177 |
},
|
178 |
"rad_gradings": {
|
179 |
"IVD label": datasets.Sequence(datasets.Value("string")),
|
@@ -319,6 +349,7 @@ class SPIDER(datasets.GeneratorBasedBuilder):
|
|
319 |
overview_dict[key] = {
|
320 |
k:v for k,v in item.items() if k not in exclude_vars
|
321 |
}
|
|
|
322 |
|
323 |
# Merge patient records for radiological gradings data
|
324 |
grades_dict = {}
|
|
|
73 |
# TODO: Add description of the dataset here
|
74 |
# You can copy an official description
|
75 |
_DESCRIPTION = """\
|
76 |
+
This is a large publicly available multi-center lumbar spine magnetic resonance \
|
77 |
+
imaging (MRI) dataset with reference segmentations of vertebrae, intervertebral \
|
78 |
+
discs (IVDs), and spinal canal. The dataset includes 447 sagittal T1 and T2 \
|
79 |
+
MRI series from 218 studies of 218 patients with a history of low back pain. \
|
80 |
+
The data was collected from four different hospitals. There is an additional \
|
81 |
+
hidden test set, not available here, used in the accompanying SPIDER challenge \
|
82 |
+
on spider.grand-challenge.org. We share this data to encourage wider \
|
83 |
+
participation and collaboration in the field of spine segmentation, and \
|
84 |
+
ultimately improve the diagnostic value of lumbar spine MRI.
|
85 |
+
|
86 |
+
This file also provides the biological sex for all patients and the age for \
|
87 |
+
the patients for which this was available. It also includes a number of \
|
88 |
+
scanner and acquisition parameters for each individual MRI study. The dataset \
|
89 |
+
also comes with radiological gradings found in a separate file for the \
|
90 |
+
following degenerative changes:
|
91 |
+
|
92 |
+
1. Modic changes (type I, II or III)
|
93 |
+
|
94 |
+
2. Upper and lower endplate changes / Schmorl nodes (binary)
|
95 |
+
|
96 |
+
3. Spondylolisthesis (binary)
|
97 |
+
|
98 |
+
4. Disc herniation (binary)
|
99 |
+
|
100 |
+
5. Disc narrowing (binary)
|
101 |
+
|
102 |
+
6. Disc bulging (binary)
|
103 |
+
|
104 |
+
7. Pfirrman grade (grade 1 to 5).
|
105 |
+
|
106 |
+
All radiological gradings are provided per IVD level."""
|
107 |
|
108 |
_HOMEPAGE = "https://zenodo.org/records/10159290"
|
109 |
|
|
|
135 |
|
136 |
|
137 |
class SPIDER(datasets.GeneratorBasedBuilder):
|
138 |
+
"""TODO: Short description of my dataset.""" #TODO
|
139 |
|
140 |
# Class attributes
|
141 |
DEFAULT_WRITER_BATCH_SIZE = 16 # PyArrow default is too large for image data
|
|
|
203 |
"TransmitCoilName": datasets.Value(dtype="string"),
|
204 |
"WindowCenter": datasets.Value(dtype="string"),
|
205 |
"WindowWidth": datasets.Value(dtype="string"),
|
206 |
+
"OrigSubset":datasets.Value(dtype="string"),
|
207 |
},
|
208 |
"rad_gradings": {
|
209 |
"IVD label": datasets.Sequence(datasets.Value("string")),
|
|
|
349 |
overview_dict[key] = {
|
350 |
k:v for k,v in item.items() if k not in exclude_vars
|
351 |
}
|
352 |
+
overview_dict[key]['OrigSubset'] = item['subset'] # Change name to original subset
|
353 |
|
354 |
# Merge patient records for radiological gradings data
|
355 |
grades_dict = {}
|