Chris Oswald
commited on
Commit
•
4b43c77
1
Parent(s):
e0f64ef
debugging
Browse files
SPIDER.py
CHANGED
@@ -142,7 +142,7 @@ class SPIDER(datasets.GeneratorBasedBuilder):
|
|
142 |
"patient_id": datasets.Value("string"),
|
143 |
"scan_type": datasets.Value("string"),
|
144 |
# "raw_image": datasets.Image(),
|
145 |
-
|
146 |
"metadata": {
|
147 |
"num_vertebrae": datasets.Value(dtype="string"),
|
148 |
"num_discs": datasets.Value(dtype="string"),
|
@@ -224,7 +224,6 @@ class SPIDER(datasets.GeneratorBasedBuilder):
|
|
224 |
# By default the archives will be extracted and a path to a cached folder where they are extracted is returned instead of the archive
|
225 |
paths_dict = dl_manager.download_and_extract(_URLS)
|
226 |
scan_types = self.config.scan_types
|
227 |
-
# scan_types = ['t1'] #TODO: remove
|
228 |
return [
|
229 |
datasets.SplitGenerator(
|
230 |
name=datasets.Split.TRAIN,
|
@@ -443,10 +442,7 @@ class SPIDER(datasets.GeneratorBasedBuilder):
|
|
443 |
# (note that only images need to be shuffled since masks and metadata
|
444 |
# will be linked to the selected image)
|
445 |
np.random.shuffle(image_files)
|
446 |
-
|
447 |
-
print(image_files)
|
448 |
-
print(mask_files)
|
449 |
-
print(len(image_files), len(mask_files))
|
450 |
## Generate next example
|
451 |
# ----------------------
|
452 |
for idx, example in enumerate(image_files):
|
@@ -463,6 +459,8 @@ class SPIDER(datasets.GeneratorBasedBuilder):
|
|
463 |
# Convert .mha image to numeric array
|
464 |
image_array = sitk.GetArrayFromImage(image)
|
465 |
|
|
|
|
|
466 |
# Extract overview data corresponding to image
|
467 |
image_overview = overview_dict[scan_id]
|
468 |
|
@@ -473,8 +471,8 @@ class SPIDER(datasets.GeneratorBasedBuilder):
|
|
473 |
return_dict = {'patient_id':patient_id, 'scan_type':scan_type}
|
474 |
# if raw_image:
|
475 |
# return_dict['raw_image'] = image
|
476 |
-
|
477 |
-
|
478 |
if metadata:
|
479 |
return_dict['metadata'] = image_overview
|
480 |
if rad_gradings:
|
|
|
142 |
"patient_id": datasets.Value("string"),
|
143 |
"scan_type": datasets.Value("string"),
|
144 |
# "raw_image": datasets.Image(),
|
145 |
+
"numeric_array": datasets.Sequence(datasets.Value("int16")),
|
146 |
"metadata": {
|
147 |
"num_vertebrae": datasets.Value(dtype="string"),
|
148 |
"num_discs": datasets.Value(dtype="string"),
|
|
|
224 |
# By default the archives will be extracted and a path to a cached folder where they are extracted is returned instead of the archive
|
225 |
paths_dict = dl_manager.download_and_extract(_URLS)
|
226 |
scan_types = self.config.scan_types
|
|
|
227 |
return [
|
228 |
datasets.SplitGenerator(
|
229 |
name=datasets.Split.TRAIN,
|
|
|
442 |
# (note that only images need to be shuffled since masks and metadata
|
443 |
# will be linked to the selected image)
|
444 |
np.random.shuffle(image_files)
|
445 |
+
|
|
|
|
|
|
|
446 |
## Generate next example
|
447 |
# ----------------------
|
448 |
for idx, example in enumerate(image_files):
|
|
|
459 |
# Convert .mha image to numeric array
|
460 |
image_array = sitk.GetArrayFromImage(image)
|
461 |
|
462 |
+
#TODO: load mask file and numeric array
|
463 |
+
|
464 |
# Extract overview data corresponding to image
|
465 |
image_overview = overview_dict[scan_id]
|
466 |
|
|
|
471 |
return_dict = {'patient_id':patient_id, 'scan_type':scan_type}
|
472 |
# if raw_image:
|
473 |
# return_dict['raw_image'] = image
|
474 |
+
if numeric_array:
|
475 |
+
return_dict['numeric_array'] = image_array
|
476 |
if metadata:
|
477 |
return_dict['metadata'] = image_overview
|
478 |
if rad_gradings:
|