def __iter__(self):
return iter([
DatasetItem(id=1, subset='train', image=np.ones((4, 4, 3)),
annotations=[
# Bbox + single polygon
Bbox(0, 1, 2, 2,
label=2, group=1, id=1,
attributes={ 'is_crowd': False }),
Polygon([0, 1, 2, 1, 2, 3, 0, 3],
attributes={ 'is_crowd': False },
label=2, group=1, id=1),
]),
DatasetItem(id=2, subset='train', image=np.ones((4, 4, 3)),
annotations=[
# Mask + bbox
Mask(np.array([
[0, 1, 0, 0],
[0, 1, 0, 0],
[0, 1, 1, 1],
[0, 0, 0, 0]],
),
attributes={ 'is_crowd': True },
label=4, group=3, id=3),
Bbox(1, 0, 2, 2, label=4, group=3, id=3,
attributes={ 'is_crowd': True }),
]),
DatasetItem(id=3, subset='val', image=np.ones((4, 4, 3)),
annotations=[
# Bbox + mask
Bbox(0, 1, 2, 2, label=4, group=3, id=3,
attributes={ 'is_crowd': True }),
Mask(np.array([
[0, 0, 0, 0],
[1, 1, 1, 0],
[1, 1, 0, 0],
[0, 0, 0, 0]],
),
attributes={ 'is_crowd': True },
label=4, group=3, id=3),
]),
])
def __iter__(self):
return iter([
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,
attributes={
'score': 1.0,
}),
Bbox(5, 6, 7, 8, id=5, group=5),
Points([1, 2, 2, 0, 1, 1], label=0, id=5),
Mask(label=3, id=5, image=np.ones((2, 3))),
]),
DatasetItem(id=21,
subset='train',
annotations=[
Caption('test'),
Label(2),
Bbox(1, 2, 3, 4, 5, id=42, group=42)
]),
DatasetItem(id=2,
subset='val',
annotations=[
PolyLine([1, 2, 3, 4, 5, 6, 7, 8], id=11),
Polygon([1, 2, 3, 4, 5, 6, 7, 8], id=12),
]),
DatasetItem(id=42, subset='test'),
DatasetItem(id=42),
DatasetItem(id=43, image=Image(path='1/b/c.qq', size=(2, 4))),
])
def test_can_save_and_load(self):
source_dataset = Dataset.from_iterable([
DatasetItem(id='1', subset='train', image=np.ones((8, 8, 3)),
annotations=[
Bbox(0, 2, 4, 2, label=0, group=1),
Points([3.2, 3.12, 4.11, 3.2, 2.11,
2.5, 3.5, 2.11, 3.8, 2.13], label=0, group=1),
]
),
DatasetItem(id='2', subset='train', image=np.ones((10, 10, 3)),
annotations=[
Points([4.23, 4.32, 5.34, 4.45, 3.54,
3.56, 4.52, 3.51, 4.78, 3.34], label=1, group=1),
]
),
DatasetItem(id='3', subset='train', image=np.ones((8, 8, 3)),
annotations=[Label(2, group=1)]
),
DatasetItem(id='4', subset='train', image=np.ones((10, 10, 3)),
annotations=[
Bbox(0, 2, 4, 2, label=3, group=1),
Points([3.2, 3.12, 4.11, 3.2, 2.11,
2.5, 3.5, 2.11, 3.8, 2.13], label=3, group=1),
]
),
DatasetItem(id='a/5', subset='train', image=np.ones((8, 8, 3)),
annotations=[
Bbox(2, 2, 2, 2, group=1),
]
),
DatasetItem(id='label_0', subset='train', image=np.ones((8, 8, 3)),
),
], categories={
AnnotationType.label: LabelCategories.from_iterable(
[('label_%s' % i, 'class_%s' % i) for i in range(5)]),
})
with TestDir() as test_dir:
VggFace2Converter.convert(source_dataset, test_dir, save_images=True)
parsed_dataset = Dataset.import_from(test_dir, 'vgg_face2')
compare_datasets(self, source_dataset, parsed_dataset)
def __iter__(self):
return iter([
DatasetItem(id='2007_000001', subset='train',
image=Image(path='2007_000001.jpg', size=(20, 10)),
annotations=[
Label(self._label(l.name))
for l in VOC.VocLabel if l.value % 2 == 1
] + [
Bbox(1, 2, 2, 2, label=self._label('cat'),
attributes={
'pose': VOC.VocPose(1).name,
'truncated': True,
'difficult': False,
'occluded': False,
},
id=1, group=1,
),
Bbox(4, 5, 2, 2, label=self._label('person'),
attributes={
'truncated': False,
'difficult': False,
'occluded': False,
**{
a.name: a.value % 2 == 1
for a in VOC.VocAction
}
},
id=2, group=2,
),
Bbox(5.5, 6, 2, 2, label=self._label(
VOC.VocBodyPart(1).name),
group=2
),
Mask(image=np.ones([5, 10]),
label=self._label(VOC.VocLabel(2).name),
group=1,
),
]
),
DatasetItem(id='2007_000002', subset='test',
image=np.zeros((20, 10, 3))),
])
def test_item_representations():
item = DatasetItem(id=1, subset='subset', path=['a', 'b'],
image=np.ones((5, 4, 3)),
annotations=[
Label(0, attributes={'a1': 1, 'a2': '2'}, id=1, group=2),
Caption('hello', id=1),
Caption('world', group=5),
Label(2, id=3, attributes={ 'x': 1, 'y': '2' }),
Bbox(1, 2, 3, 4, label=4, id=4, attributes={ 'a': 1.0 }),
Bbox(5, 6, 7, 8, id=5, group=5),
Points([1, 2, 2, 0, 1, 1], label=0, id=5),
Mask(id=5, image=np.ones((3, 2))),
Mask(label=3, id=5, image=np.ones((2, 3))),
PolyLine([1, 2, 3, 4, 5, 6, 7, 8], id=11),
Polygon([1, 2, 3, 4, 5, 6, 7, 8]),
]
)
encoded = DatasetItemEncoder.encode(item)
DatasetItemEncoder.to_string(encoded)
def test_can_import(self):
expected_dataset = Dataset.from_iterable(
[
DatasetItem(id=1,
subset='train',
image=np.ones((10, 15, 3)),
annotations=[
Bbox(0, 2, 4, 2, label=2),
Bbox(3, 3, 2, 3, label=4),
]),
],
categories={
AnnotationType.label:
LabelCategories.from_iterable('label_' + str(i)
for i in range(10)),
})
dataset = Dataset.import_from(DUMMY_DATASET_DIR, 'yolo')
compare_datasets(self, expected_dataset, dataset)
def __iter__(self):
return iter([
DatasetItem(id=1, subset='train', annotations=[
Bbox(1, 2, 3, 4),
Label(4),
]),
DatasetItem(id=1, subset='val', annotations=[
Label(4),
Label(5),
]),
])
def test_can_save_and_load_with_no_subsets(self):
expected_dataset = Dataset.from_iterable([
DatasetItem(id=1, image=np.ones((8, 8, 3)),
annotations=[
Bbox(0, 1, 3, 5),
]),
])
with TestDir() as test_dir:
self._test_save_and_load(expected_dataset,
IcdarTextLocalizationConverter.convert, test_dir)
def test_can_save_bboxes(self):
test_dataset = Dataset.from_iterable([
DatasetItem(id=1, subset='train',
image=np.ones((16, 16, 3)),
annotations=[
Bbox(0, 4, 4, 8, label=2),
Bbox(0, 4, 4, 4, label=3),
Bbox(2, 4, 4, 4),
], attributes={'source_id': ''}
),
], categories={
AnnotationType.label: LabelCategories.from_iterable(
'label_' + str(label) for label in range(10)),
})
with TestDir() as test_dir:
self._test_save_and_load(
test_dataset,
partial(TfDetectionApiConverter.convert, save_images=True),
test_dir)
def __iter__(self):
yield DatasetItem(id=1, annotations=[
# drop non voc label
Bbox(2, 3, 4, 5, label=self._label('cat'), id=1, group=1,
attributes={
'truncated': False,
'difficult': False,
'occluded': False,
}
),
])
def __iter__(self):
yield DatasetItem(id=1,
annotations=[
Bbox(1,
2,
3,
4,
label=0,
id=1,
group=1,
attributes={
'act1': True,
'act2': False,
'truncated': False,
'difficult': False,
'occluded': False,
}),
Bbox(2, 3, 4, 5, label=1, group=1),
Bbox(2, 3, 4, 6, label=2, group=1),
])
def __iter__(self):
return iter([
DatasetItem(id=0,
subset='s1',
image=np.zeros((5, 10, 3)),
annotations=[
Polygon([0, 0, 4, 0, 4, 4],
label=1,
group=4,
attributes={'occluded': True}),
Polygon([5, 0, 9, 0, 5, 5],
label=2,
group=4,
attributes={'unknown': 'bar'}),
Points([1, 1, 3, 2, 2, 3],
label=2,
attributes={
'a1': 'x',
'a2': 42
}),
Label(1),
Label(2, attributes={
'a1': 'y',
'a2': 44
}),
]),
DatasetItem(id=1,
subset='s1',
annotations=[
PolyLine([0, 0, 4, 0, 4, 4],
label=3,
id=4,
group=4),
Bbox(5, 0, 1, 9, label=3, id=4, group=4),
]),
DatasetItem(
id=2,
subset='s2',
image=np.ones((5, 10, 3)),
annotations=[
Polygon([0, 0, 4, 0, 4, 4],
label=3,
group=4,
attributes={
'z_order': 1,
'occluded': False
}),
PolyLine([5, 0, 9, 0, 5,
5]), # will be skipped as no label
]),
DatasetItem(id=3,
subset='s3',
image=Image(path='3.jpg', size=(2, 4))),
])
def __iter__(self):
return iter([
DatasetItem(id=1,
annotations=[
Label(1),
Bbox(1, 2, 3, 4, label=0),
Mask(image=np.array([1]), label=1),
Polygon([1, 1, 2, 2, 3, 4], label=2),
PolyLine([1, 3, 4, 2, 5, 6], label=None)
]),
])
def test_can_load_dataset_with_exact_image_info(self):
source_dataset = Dataset.from_iterable([
DatasetItem(id=1, subset='train',
image=Image(path='1.jpg', size=(10, 15)),
annotations=[
Bbox(0, 2, 4, 2, label=2),
Bbox(3, 3, 2, 3, label=4),
]),
], categories={
AnnotationType.label: LabelCategories.from_iterable(
'label_' + str(i) for i in range(10)),
})
with TestDir() as test_dir:
YoloConverter.convert(source_dataset, test_dir)
parsed_dataset = YoloImporter()(test_dir,
image_info={'1': (10, 15)}).make_dataset()
compare_datasets(self, source_dataset, parsed_dataset)
def setUpClass(cls):
cls.dataset = Dataset.from_iterable([
DatasetItem(id=1, image=np.ones((5, 5, 3)), annotations=[
Label(1, id=0, attributes={'a': 1, 'b': 7, }),
Bbox(1, 2, 3, 4, id=1, label=0, attributes={
'a': 1, 'b': 2,
}),
]),
DatasetItem(id=2, image=np.ones((2, 4, 3)), annotations=[
Label(2, id=0, attributes={'a': 2, 'b': 2, }),
Bbox(2, 3, 1, 4, id=1, label=0, attributes={
'a': 1, 'b': 1,
}),
]),
DatasetItem(id=3),
DatasetItem(id=4, image=np.ones((2, 4, 3)), annotations=[
Label(0, id=0, attributes={'b': 4, }),
Label(1, id=1, attributes={'a': 11, 'b': 7, }),
Bbox(1, 3, 2, 4, id=2, label=0, attributes={
'a': 2, 'b': 1,
}),
Bbox(3, 1, 4, 2, id=3, label=0, attributes={
'a': 2, 'b': 2,
}),
]),
DatasetItem(id=5, image=np.ones((2, 4, 3)), annotations=[
Label(0, id=0, attributes={'a': 20, 'b': 10, }),
Bbox(1, 2, 3, 4, id=1, label=1, attributes={
'a': 1, 'b': 1,
}),
]),
DatasetItem(id=6, image=np.ones((2, 4, 3)), annotations=[
Label(1, id=0, attributes={'a': 11, 'b': 2, 'c': 3, }),
Bbox(2, 3, 4, 1, id=1, label=1, attributes={
'a': 2, 'b': 2,
}),
]),
DatasetItem(id=7, image=np.ones((2, 4, 3)), annotations=[
Label(1, id=0, attributes={'a': 1, 'b': 2, 'c': 5, }),
Bbox(1, 2, 3, 4, id=1, label=2, attributes={
'a': 1, 'b': 2,
}),
]),
DatasetItem(id=8, image=np.ones((2, 4, 3)), annotations=[
Label(2, id=0, attributes={'a': 7, 'b': 9, 'c': 5, }),
Bbox(2, 1, 3, 4, id=1, label=2, attributes={
'a': 2, 'b': 1,
}),
]),
], categories=[[f'label_{i}', None, {'a', 'b', }]
for i in range(2)])
def __iter__(self):
return iter([
DatasetItem(id=1, subset='train',
image=np.ones((16, 16, 3)),
annotations=[
Bbox(0, 4, 4, 8, label=2, group=2),
Polygon([0, 4, 4, 4, 5, 6], label=3, attributes={
'occluded': True,
'a1': 'qwe',
'a2': True,
'a3': 123,
}),
Mask(np.array([[0, 1], [1, 0], [1, 1]]), group=2,
attributes={ 'username': '******' }),
Bbox(1, 2, 3, 4, group=3),
Mask(np.array([[0, 0], [0, 0], [1, 1]]), group=3,
attributes={ 'occluded': True }
),
]
),
])
def append_bbox_mot(annotations, **kwargs):
annotations.append(
Bbox(
1,
1,
2,
2,
label=kwargs["label_id"],
attributes=kwargs["attributes"],
))
annotations.append(
Label(kwargs["label_id"], attributes=kwargs["attributes"]))
def test_can_save_dataset_with_no_labels(self):
source_dataset = Dataset.from_iterable([
DatasetItem(id='1', image=np.ones((8, 8, 3)),
annotations=[
Bbox(0, 2, 4, 2, group=1),
Points([4.23, 4.32, 5.34, 4.45, 3.54,
3.56, 4.52, 3.51, 4.78, 3.34], group=1),
]
),
DatasetItem(id='2', image=np.ones((8, 8, 3)),
annotations=[
Bbox(2, 2, 4, 2, group=1),
]
),
], categories=[])
with TestDir() as test_dir:
VggFace2Converter.convert(source_dataset, test_dir, save_images=False)
parsed_dataset = Dataset.import_from(test_dir, 'vgg_face2')
compare_datasets(self, source_dataset, parsed_dataset)
def test_can_save_dataset_with_image_info(self):
source_dataset = Dataset.from_iterable([
DatasetItem(id=1, subset='train',
image=Image(path='1.jpg', size=(10, 15)),
annotations=[
Bbox(0, 2, 4, 2, label=2),
Bbox(3, 3, 2, 3, label=4),
]),
], categories={
AnnotationType.label: LabelCategories.from_iterable(
'label_' + str(i) for i in range(10)),
})
with TestDir() as test_dir:
YoloConverter.convert(source_dataset, test_dir)
save_image(osp.join(test_dir, 'obj_train_data', '1.jpg'),
np.ones((10, 15, 3))) # put the image for dataset
parsed_dataset = YoloImporter()(test_dir).make_dataset()
compare_datasets(self, source_dataset, parsed_dataset)
def __iter__(self):
return iter([
DatasetItem(id=1,
subset='train',
image=np.ones((16, 16, 3)),
annotations=[
Bbox(0,
4,
4,
8,
label=2,
attributes={
'occluded': True,
}),
Bbox(0,
4,
4,
4,
label=3,
attributes={
'visibility': 0.4,
}),
Bbox(2,
4,
4,
4,
attributes={'ignored': True}),
]),
DatasetItem(id=2,
subset='val',
image=np.ones((8, 8, 3)),
annotations=[
Bbox(1, 2, 4, 2, label=3),
]),
DatasetItem(
id=3,
subset='test',
image=np.ones((5, 4, 3)) * 3,
),
])
def test_can_import_bboxes(self):
expected_dataset = Dataset.from_iterable([
DatasetItem(id='img_1', subset='train',
image=np.ones((10, 15, 3)),
annotations=[
Polygon([0, 0, 3, 1, 4, 6, 1, 7],
attributes={'text': 'FOOD'}),
]
),
DatasetItem(id='img_2', subset='train',
image=np.ones((10, 15, 3)),
annotations=[
Bbox(0, 0, 2, 3, attributes={'text': 'RED'}),
Bbox(3, 3, 2, 3, attributes={'text': 'LION'}),
]
),
])
dataset = Dataset.import_from(
osp.join(DUMMY_DATASET_DIR, 'text_localization'), 'icdar')
compare_datasets(self, expected_dataset, dataset)
def test_can_find_missing_boxes(self):
detections = 3
class_count = 2
item1 = DatasetItem(id=1, annotations=[
Bbox(i * 10, 10, 10, 10, label=i)
for i in range(detections) if i % 2 == 0
])
item2 = DatasetItem(id=2, annotations=[
Bbox(i * 10, 10, 10, 10, label=(i + 1) % class_count)
for i in range(detections) if i % 2 == 1
])
iou_thresh = 0.5
comp = DistanceComparator(iou_threshold=iou_thresh)
result = comp.match_boxes(item1, item2)
matches, mispred, a_greater, b_greater = result
self.assertEqual(0, len(mispred))
self.assertEqual(len(item1.annotations), len(a_greater))
self.assertEqual(len(item2.annotations), len(b_greater))
self.assertEqual(0, len(matches))
def __iter__(self):
return iter([
DatasetItem(id='a', annotations=[
Bbox(2, 3, 4, 5, label=2, id=1, group=1,
attributes={
'truncated': False,
'difficult': False,
'occluded': True,
'x': '1', 'y': '2', # can only read strings
}
),
]),
])
def __iter__(self):
return iter([
DatasetItem(id=1,
subset='train',
image=np.ones((16, 16, 3)),
annotations=[
Bbox(0, 4, 4, 8, label=2),
Bbox(0, 4, 4, 4, label=3),
Bbox(2, 4, 4, 4),
]),
DatasetItem(id=2,
subset='val',
image=np.ones((8, 8, 3)),
annotations=[
Bbox(1, 2, 4, 2, label=3),
]),
DatasetItem(
id=3,
subset='test',
image=np.ones((5, 4, 3)) * 3,
),
])
def __iter__(self):
return iter([
DatasetItem(id=1, annotations=[
# Should be remapped
Label(1),
Bbox(1, 2, 3, 4, label=2),
Mask(image=np.array([1]), label=3),
# Should be kept
Polygon([1, 1, 2, 2, 3, 4], label=4),
PolyLine([1, 3, 4, 2, 5, 6], label=None)
]),
])
def test_can_merge_classes(self):
source0 = Dataset.from_iterable([
DatasetItem(1,
annotations=[
Label(0),
Label(1),
Bbox(0, 0, 1, 1, label=1),
]),
],
categories=['a', 'b'])
source1 = Dataset.from_iterable([
DatasetItem(1,
annotations=[
Label(0),
Label(1),
Bbox(0, 0, 1, 1, label=0),
Bbox(0, 0, 1, 1, label=1),
]),
],
categories=['b', 'c'])
expected = Dataset.from_iterable([
DatasetItem(1,
annotations=[
Label(0),
Label(1),
Label(2),
Bbox(0, 0, 1, 1, label=1),
Bbox(0, 0, 1, 1, label=2),
]),
],
categories=['a', 'b', 'c'])
merger = IntersectMerge()
merged = merger([source0, source1])
compare_datasets(self, expected, merged, ignored_attrs={'score'})
def append_bbox_coco(annotations, **kwargs):
annotations.append(
Bbox(
1,
1,
2,
2,
label=kwargs["label_id"],
id=kwargs["ann_id"],
attributes=kwargs["attributes"],
group=kwargs["ann_id"],
))
annotations.append(
Label(kwargs["label_id"], attributes=kwargs["attributes"]))
def _process(self, image):
detections = []
for i, roi in enumerate(self.rois):
roi_sum = self.roi_value(roi, image)
roi_base_sum = self.roi_base_sums[i]
first_run = roi_base_sum is None
if first_run:
roi_base_sum = roi_sum
self.roi_base_sums[i] = roi_base_sum
cls_conf = roi_sum / roi_base_sum
if roi.threshold < roi_sum / roi_base_sum:
cls = roi.label
detections.append(
Bbox(roi.x,
roi.y,
roi.w,
roi.h,
label=cls,
attributes={'score': cls_conf}))
if first_run:
continue
for j in range(self.fp_count):
if roi.threshold < cls_conf:
cls = roi.label
else:
cls = (i + j) % self.class_count
box = [roi.x, roi.y, roi.w, roi.h]
offset = (np.random.rand(4) - 0.5) * self.pixel_jitter
detections.append(
Bbox(*(box + offset),
label=cls,
attributes={'score': cls_conf}))
return detections
def download_labeled_image(self, task_id: int, frame_id: int, outdir: str):
filename, img = self.task.download_frame(task_id,
frame_id,
outdir,
save=False)
labels = dict([(label.id, label)
for label in self.task.get_id(task_id).labels])
jobs = self.task.get_jobs(task_id)
job_id_with_frame = [
job.id for job in jobs if (int(job.start_frame) <= frame_id)
& (frame_id <= int(job.stop_frame))
]
job_annos_list = [
self.job.get_annotations(job_id) for job_id in job_id_with_frame
]
annos = [[anno for anno in job_annos.shapes
if anno.frame == frame_id][0]
for job_annos in job_annos_list]
categories = [label.name for label in labels.values()]
items = []
for anno in annos:
attriname = dict([(attr.id, attr.name)
for attr in labels[anno.label_id].attributes])
attris = dict([(attriname[attr.spec_id], attr.value)
for attr in anno.attributes])
if anno.type == 'rectangle':
x, y, x2, y2 = anno.points
items.append(
Bbox(x,
y,
x2 - x,
y2 - y,
attributes=attris,
label=categories.index(labels[anno.label_id].name)))
elif anno.type == 'polygon':
items.append(
Polygon(anno.points,
attributes=attris,
label=categories.index(
labels[anno.label_id].name)))
# image = {'data': np.asarray(img),
# 'path': str(Path(outdir)/filename)}
dataset = Dataset.from_iterable(
[DatasetItem(id=filename, annotations=items, image=np.array(img))],
categories=categories)
customset = customDataset(dataset)
for image_data in customset._imageDatas:
image_data.drawItem('s', 's').saveImg(Path(outdir) / filename)
def test_can_save_and_load_bboxes(self):
expected_dataset = Dataset.from_iterable([
DatasetItem(id=1, subset='train',
annotations=[
Bbox(1, 3, 6, 10),
Bbox(0, 1, 3, 5, attributes={'text': 'word_0'}),
]),
DatasetItem(id=2, subset='train',
annotations=[
Polygon([0, 0, 3, 0, 4, 7, 1, 8],
attributes={'text': 'word_1'}),
Polygon([1, 2, 5, 3, 6, 8, 0, 7]),
]),
DatasetItem(id=3, subset='train',
annotations=[
Polygon([2, 2, 8, 3, 7, 10, 2, 9],
attributes={'text': 'word_2'}),
Bbox(0, 2, 5, 9, attributes={'text': 'word_3'}),
]),
])
with TestDir() as test_dir:
self._test_save_and_load(expected_dataset,
IcdarTextLocalizationConverter.convert, test_dir)
