def categories(self):
label_cat = LabelCategories()
for label in range(10):
label_cat.add('label_' + str(label))
return {
AnnotationType.label: label_cat,
}
def _generate_classification_dataset(self, config, num_duplicate):
subsets = ["train", "val", "test"]
dummy_images = [
np.random.randint(0, 255, size=(224, 224, 3))
for _ in range(num_duplicate)
]
iterable = []
label_cat = LabelCategories()
idx = 0
for label_id, label in enumerate(config.keys()):
label_cat.add(label, attributes=None)
num_item = config[label]
for subset in subsets:
for _ in range(num_item):
idx += 1
iterable.append(
DatasetItem(
idx,
subset=subset,
annotations=[Label(label_id)],
image=Image(data=dummy_images[idx %
num_duplicate]),
))
categories = {AnnotationType.label: label_cat}
dataset = Dataset.from_iterable(iterable, categories)
return dataset
def categories(self):
label_cat = LabelCategories()
for label in VOC.VocLabel:
label_cat.add(label.name)
return {
AnnotationType.label: label_cat,
}
def categories(self):
label_categories = LabelCategories()
for i in range(10):
label_categories.add(str(i))
return {
AnnotationType.label: label_categories,
}
def categories(self):
label_categories = LabelCategories()
label_categories.add('label1', attributes={'a1', 'a2'})
label_categories.add('label2')
return {
AnnotationType.label: label_categories,
}
def categories(self):
label_cat = LabelCategories()
label_cat.add('Label_1') # should become lowercase
label_cat.add('label_2')
return {
AnnotationType.label: label_cat,
}
def _make_label_id_map(self, src_label_cat, label_mapping, default_action):
dst_label_cat = LabelCategories(attributes=src_label_cat.attributes)
id_mapping = {}
for src_index, src_label in enumerate(src_label_cat.items):
dst_label = label_mapping.get(src_label.name)
if not dst_label and default_action == self.DefaultAction.keep:
dst_label = src_label.name # keep unspecified as is
if not dst_label:
continue
dst_index = dst_label_cat.find(dst_label)[0]
if dst_index is None:
dst_label_cat.add(dst_label,
src_label.parent, src_label.attributes)
dst_index = dst_label_cat.find(dst_label)[0]
id_mapping[src_index] = dst_index
if log.getLogger().isEnabledFor(log.DEBUG):
log.debug("Label mapping:")
for src_id, src_label in enumerate(src_label_cat.items):
if id_mapping.get(src_id):
log.debug("#%s '%s' -> #%s '%s'",
src_id, src_label.name, id_mapping[src_id],
dst_label_cat.items[id_mapping[src_id]].name
)
else:
log.debug("#%s '%s' -> <deleted>", src_id, src_label.name)
self._map_id = lambda src_id: id_mapping.get(src_id, None)
self._categories[AnnotationType.label] = dst_label_cat
def make_camvid_categories(label_map=None):
if label_map is None:
label_map = CamvidLabelMap
# There must always be a label with color (0, 0, 0) at index 0
bg_label = find(label_map.items(), lambda x: x[1] == (0, 0, 0))
if bg_label is not None:
bg_label = bg_label[0]
else:
bg_label = 'background'
if bg_label not in label_map:
has_colors = any(v is not None for v in label_map.values())
color = (0, 0, 0) if has_colors else None
label_map[bg_label] = color
label_map.move_to_end(bg_label, last=False)
categories = {}
label_categories = LabelCategories()
for label, desc in label_map.items():
label_categories.add(label)
categories[AnnotationType.label] = label_categories
has_colors = any(v is not None for v in label_map.values())
if not has_colors: # generate new colors
colormap = generate_colormap(len(label_map))
else: # only copy defined colors
label_id = lambda label: label_categories.find(label)[0]
colormap = {
label_id(name): (desc[0], desc[1], desc[2])
for name, desc in label_map.items()
}
mask_categories = MaskCategories(colormap)
mask_categories.inverse_colormap # pylint: disable=pointless-statement
categories[AnnotationType.mask] = mask_categories
return categories
def _load_categories(parsed):
categories = {}
parsed_label_cat = parsed['categories'].get(AnnotationType.label.name)
if parsed_label_cat:
label_categories = LabelCategories()
for item in parsed_label_cat['labels']:
label_categories.add(item['name'], parent=item['parent'])
categories[AnnotationType.label] = label_categories
parsed_mask_cat = parsed['categories'].get(AnnotationType.mask.name)
if parsed_mask_cat:
colormap = {}
for item in parsed_mask_cat['colormap']:
colormap[int(item['label_id'])] = \
(item['r'], item['g'], item['b'])
mask_categories = MaskCategories(colormap=colormap)
categories[AnnotationType.mask] = mask_categories
parsed_points_cat = parsed['categories'].get(
AnnotationType.points.name)
if parsed_points_cat:
point_categories = PointsCategories()
for item in parsed_points_cat['items']:
point_categories.add(int(item['label_id']),
item['labels'],
joints=item['joints'])
categories[AnnotationType.points] = point_categories
return categories
def categories(self):
label_cat = LabelCategories()
label_cat.add(VOC.VocLabel.cat.name)
label_cat.add('non_voc_label')
return {
AnnotationType.label: label_cat,
}
def make_voc_categories(label_map=None):
if label_map is None:
label_map = make_voc_label_map()
categories = {}
label_categories = LabelCategories()
label_categories.attributes.update(['difficult', 'truncated', 'occluded'])
for label, desc in label_map.items():
label_categories.add(label, attributes=desc[2])
for part in OrderedDict((k, None) for k in chain(
*(desc[1] for desc in label_map.values()))):
label_categories.add(part)
categories[AnnotationType.label] = label_categories
has_colors = any(v[0] is not None for v in label_map.values())
if not has_colors: # generate new colors
colormap = generate_colormap(len(label_map))
else: # only copy defined colors
label_id = lambda label: label_categories.find(label)[0]
colormap = { label_id(name): desc[0]
for name, desc in label_map.items() if desc[0] is not None }
mask_categories = MaskCategories(colormap)
mask_categories.inverse_colormap # pylint: disable=pointless-statement
categories[AnnotationType.mask] = mask_categories
return categories
def categories(self):
label_cat = LabelCategories()
label_cat.add('Label_1')
label_cat.add('label_2')
return {
AnnotationType.label: label_cat,
}
def test_can_convert_polygons_to_mask(self):
label_categories = LabelCategories()
for i in range(10):
label_categories.add(str(i))
class SrcTestExtractor(Extractor):
def __iter__(self):
return iter([
DatasetItem(id=1,
image=np.zeros((6, 10, 3)),
annotations=[
Polygon([0, 0, 4, 0, 4, 4],
label=3,
id=4,
group=4),
Polygon([5, 0, 9, 0, 5, 5],
label=3,
id=4,
group=4),
]),
])
def categories(self):
return {AnnotationType.label: label_categories}
class DstTestExtractor(Extractor):
def __iter__(self):
return iter([
DatasetItem(
id=1,
image=np.zeros((6, 10, 3)),
annotations=[
Mask(
np.array(
[[0, 1, 1, 1, 0, 1, 1, 1, 1, 0],
[0, 0, 1, 1, 0, 1, 1, 1, 0, 0],
[0, 0, 0, 1, 0, 1, 1, 0, 0, 0],
[0, 0, 0, 0, 0, 1, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0]],
# only internal fragment (without the border),
# but not everywhere...
),
attributes={'is_crowd': True},
label=3,
id=4,
group=4),
],
attributes={'id': 1}),
])
def categories(self):
return {AnnotationType.label: label_categories}
with TestDir() as test_dir:
self._test_save_and_load(
SrcTestExtractor(),
CocoInstancesConverter(segmentation_mode='mask'),
test_dir,
target_dataset=DstTestExtractor())
def test_can_crop_covered_segments(self):
label_categories = LabelCategories()
for i in range(10):
label_categories.add(str(i))
class SrcTestExtractor(Extractor):
def __iter__(self):
return iter([
DatasetItem(id=1,
image=np.zeros((5, 5, 3)),
annotations=[
Mask(np.array(
[[0, 0, 1, 1, 1], [0, 0, 1, 1, 1],
[1, 1, 0, 1, 1], [1, 1, 1, 0, 0],
[1, 1, 1, 0, 0]], ),
label=2,
id=1,
z_order=0),
Polygon([1, 1, 4, 1, 4, 4, 1, 4],
label=1,
id=2,
z_order=1),
]),
])
def categories(self):
return {AnnotationType.label: label_categories}
class DstTestExtractor(Extractor):
def __iter__(self):
return iter([
DatasetItem(id=1,
image=np.zeros((5, 5, 3)),
annotations=[
Mask(np.array(
[[0, 0, 1, 1, 1], [0, 0, 0, 0, 1],
[1, 0, 0, 0, 1], [1, 0, 0, 0, 0],
[1, 1, 1, 0, 0]], ),
attributes={'is_crowd': True},
label=2,
id=1,
group=1),
Polygon([1, 1, 4, 1, 4, 4, 1, 4],
label=1,
id=2,
group=2,
attributes={'is_crowd': False}),
],
attributes={'id': 1}),
])
def categories(self):
return {AnnotationType.label: label_categories}
with TestDir() as test_dir:
self._test_save_and_load(SrcTestExtractor(),
CocoInstancesConverter(crop_covered=True),
test_dir,
target_dataset=DstTestExtractor())
def _load_categories(names_path):
label_categories = LabelCategories()
with open(names_path, 'r') as f:
for label in f:
label_categories.add(label.strip())
return label_categories
def test_can_convert_masks_to_polygons(self):
label_categories = LabelCategories()
for i in range(10):
label_categories.add(str(i))
class SrcExtractor(Extractor):
def __iter__(self):
return iter([
DatasetItem(id=1,
image=np.zeros((5, 10, 3)),
annotations=[
Mask(np.array([
[0, 1, 1, 1, 0, 1, 1, 1, 1, 0],
[0, 0, 1, 1, 0, 1, 1, 1, 0, 0],
[0, 0, 0, 1, 0, 1, 1, 0, 0, 0],
[0, 0, 0, 0, 0, 1, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
]),
label=3,
id=4,
group=4),
]),
])
def categories(self):
return {AnnotationType.label: label_categories}
class DstExtractor(Extractor):
def __iter__(self):
return iter([
DatasetItem(
id=1,
image=np.zeros((5, 10, 3)),
annotations=[
Polygon([3.0, 2.5, 1.0, 0.0, 3.5, 0.0, 3.0, 2.5],
label=3,
id=4,
group=4,
attributes={'is_crowd': False}),
Polygon([5.0, 3.5, 4.5, 0.0, 8.0, 0.0, 5.0, 3.5],
label=3,
id=4,
group=4,
attributes={'is_crowd': False}),
],
attributes={'id': 1}),
])
def categories(self):
return {AnnotationType.label: label_categories}
with TestDir() as test_dir:
self._test_save_and_load(
SrcExtractor(),
CocoInstancesConverter(segmentation_mode='polygons'),
test_dir,
target_dataset=DstExtractor())
def test_can_merge_polygons_on_loading(self):
label_categories = LabelCategories()
for i in range(10):
label_categories.add(str(i))
categories = {AnnotationType.label: label_categories}
class TestExtractor(Extractor):
def __iter__(self):
return iter([
DatasetItem(id=1,
image=np.zeros((6, 10, 3)),
annotations=[
Polygon([0, 0, 4, 0, 4, 4],
label=3,
id=4,
group=4),
Polygon([5, 0, 9, 0, 5, 5],
label=3,
id=4,
group=4),
]),
])
def categories(self):
return categories
class TargetExtractor(TestExtractor):
def __iter__(self):
items = list(super().__iter__())
items[0]._annotations = [
Mask(
np.array(
[[0, 1, 1, 1, 0, 1, 1, 1, 1, 0],
[0, 0, 1, 1, 0, 1, 1, 1, 0, 0],
[0, 0, 0, 1, 0, 1, 1, 0, 0, 0],
[0, 0, 0, 0, 0, 1, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0]],
# only internal fragment (without the border),
# but not everywhere...
),
label=3,
id=4,
group=4,
attributes={'is_crowd': False}),
]
return iter(items)
with TestDir() as test_dir:
self._test_save_and_load(
TestExtractor(),
CocoInstancesConverter(),
test_dir,
importer_args={'merge_instance_polygons': True},
target_dataset=TargetExtractor())
def categories(self):
label_categories = LabelCategories()
points_categories = PointsCategories()
for i in range(10):
label_categories.add(str(i))
points_categories.add(i, [])
return {
AnnotationType.label: label_categories,
AnnotationType.points: points_categories,
}
def get_categories():
# output categories - label map etc.
label_categories = LabelCategories()
with open("samples/imagenet.class", "r") as file:
for line in file.readlines():
label = line.strip()
label_categories.add(label)
return {AnnotationType.label: label_categories}
def _init_categories(self):
if self._adapter is None or self._adapter.label_map is None:
return None
label_map = sorted(self._adapter.label_map.items(), key=lambda e: e[0])
label_cat = LabelCategories()
for _, label in label_map:
label_cat.add(label)
return { AnnotationType.label: label_cat }
def _load_categories(self, labels):
attributes = ['track_id']
if self._is_gt:
attributes += ['occluded', 'visibility', 'ignored']
else:
attributes += ['score']
label_cat = LabelCategories(attributes=attributes)
for label in labels:
label_cat.add(label)
return { AnnotationType.label: label_cat }
def _load_label_categories(self, loader):
catIds = loader.getCatIds()
cats = loader.loadCats(catIds)
categories = LabelCategories()
label_map = {}
for idx, cat in enumerate(cats):
label_map[cat['id']] = idx
categories.add(name=cat['name'], parent=cat['supercategory'])
return categories, label_map
def categories(self):
label_cat = LabelCategories()
label_cat.add('window')
label_cat.add('license plate')
label_cat.add('o1')
label_cat.add('q1')
label_cat.add('b1')
label_cat.add('m1')
label_cat.add('hg')
return {
AnnotationType.label: label_cat,
}
def test_dataset(self):
label_categories = LabelCategories()
for i in range(5):
label_categories.add('cat' + str(i))
mask_categories = MaskCategories(
generate_colormap(len(label_categories.items)))
points_categories = PointsCategories()
for index, _ in enumerate(label_categories.items):
points_categories.add(index, ['cat1', 'cat2'], joints=[[0, 1]])
return Dataset.from_iterable([
DatasetItem(id=100, subset='train', image=np.ones((10, 6, 3)),
annotations=[
Caption('hello', id=1),
Caption('world', id=2, group=5),
Label(2, id=3, attributes={
'x': 1,
'y': '2',
}),
Bbox(1, 2, 3, 4, label=4, id=4, z_order=1, attributes={
'score': 1.0,
}),
Bbox(5, 6, 7, 8, id=5, group=5),
Points([1, 2, 2, 0, 1, 1], label=0, id=5, z_order=4),
Mask(label=3, id=5, z_order=2, image=np.ones((2, 3))),
]),
DatasetItem(id=21, subset='train',
annotations=[
Caption('test'),
Label(2),
Bbox(1, 2, 3, 4, label=5, id=42, group=42)
]),
DatasetItem(id=2, subset='val',
annotations=[
PolyLine([1, 2, 3, 4, 5, 6, 7, 8], id=11, z_order=1),
Polygon([1, 2, 3, 4, 5, 6, 7, 8], id=12, z_order=4),
]),
DatasetItem(id=42, subset='test',
attributes={'a1': 5, 'a2': '42'}),
DatasetItem(id=42),
DatasetItem(id=43, image=Image(path='1/b/c.qq', size=(2, 4))),
], categories={
AnnotationType.label: label_categories,
AnnotationType.mask: mask_categories,
AnnotationType.points: points_categories,
})
def categories(self):
label_cat = LabelCategories()
label_cat.add('label0')
label_cat.add('label9')
label_cat.add('label4')
mask_cat = MaskCategories(colormap={
k: v for k, v in mask_tools.generate_colormap(5).items()
if k in { 0, 1, 3, 4 }
})
return {
AnnotationType.label: label_cat,
AnnotationType.mask: mask_cat,
}
def test_can_allow_undeclared_attrs(self):
source_dataset = Dataset.from_iterable([
DatasetItem(
id=0,
annotations=[
Label(0, attributes={
'x': 4,
'y': 2
}),
Bbox(1, 2, 3, 4, label=0, attributes={
'x': 1,
'y': 1
}),
]),
],
categories=[('a', '', {'x'})])
target_label_cat = LabelCategories(attributes={'occluded'})
target_label_cat.add('a', attributes={'x'})
target_dataset = Dataset.from_iterable(
[
DatasetItem(id=0,
annotations=[
Label(0, attributes={
'x': 4,
'y': 2
}),
Bbox(1,
2,
3,
4,
label=0,
attributes={
'x': 1,
'y': 1,
'occluded': False
}),
],
attributes={'frame': 0}),
],
categories={AnnotationType.label: target_label_cat})
with TestDir() as test_dir:
self._test_save_and_load(source_dataset,
partial(CvatConverter.convert,
allow_undeclared_attrs=True),
test_dir,
target_dataset=target_dataset)
def categories(self):
label_categories = LabelCategories()
for i in range(5):
label_categories.add('cat' + str(i))
mask_categories = MaskCategories(
generate_colormap(len(label_categories.items)))
points_categories = PointsCategories()
for index, _ in enumerate(label_categories.items):
points_categories.add(index, ['cat1', 'cat2'], adjacent=[0, 1])
return {
AnnotationType.label: label_categories,
AnnotationType.mask: mask_categories,
AnnotationType.points: points_categories,
}
def _make_voc_categories():
categories = {}
label_categories = LabelCategories()
for label in chain(VocLabel, VocAction, VocBodyPart):
label_categories.add(label.name)
categories[AnnotationType.label] = label_categories
def label_id(class_index):
class_label = VocLabel(class_index).name
label_id, _ = label_categories.find(class_label)
return label_id
colormap = { label_id(idx): tuple(color) \
for idx, color in VocColormap.items() }
mask_categories = MaskCategories(colormap)
mask_categories.inverse_colormap # init inverse colormap
categories[AnnotationType.mask] = mask_categories
return categories
def _generate_classification_dataset(self,
config,
subset=None,
empty_scores=False,
out_range=False,
no_attr=False,
no_img=False):
probs = self._get_probs(out_range)
if subset is None:
self.subset = ["train", "val", "test"]
else:
self.subset = subset
iterable = []
label_cat = LabelCategories()
idx = 0
for label_id, label in enumerate(config.keys()):
num_item = config[label]
label_cat.add(label, attributes=None)
for _ in range(num_item):
scores = probs[idx]
idx += 1
if empty_scores:
scores = []
attr = {"scores": scores}
if no_attr:
attr = {}
img = Image(path=f"test/dataset/{idx}.jpg", size=(90, 90))
if no_img:
img = None
iterable.append(
DatasetItem(
idx,
subset=self.subset[idx % len(self.subset)],
annotations=[Label(
label_id,
attributes=attr,
)],
image=img,
))
categories = {AnnotationType.label: label_cat}
dataset = Dataset.from_iterable(iterable, categories)
return dataset
def __init__(self, extractor, label):
super().__init__(extractor)
assert isinstance(label, str)
self._categories = {}
label_cat = self._extractor.categories().get(AnnotationType.label)
if not label_cat:
label_cat = LabelCategories()
self._label = label_cat.add(label)
self._categories[AnnotationType.label] = label_cat