def _make_image_labels_dataset(img,
images_dir,
num_samples=4,
num_objects_per_sample=3):
exts = [".jpg", ".png"]
samples = []
for idx in range(num_samples):
filepath = os.path.join(images_dir,
"%06d%s" % (idx, exts[idx % len(exts)]))
etai.write(img, filepath)
image_labels = etai.ImageLabels()
_label = random.choice(["sun", "rain", "snow"])
image_labels.add_attribute(etad.CategoricalAttribute("label", _label))
for _ in range(num_objects_per_sample):
_label = random.choice(["cat", "dog", "bird", "rabbit"])
_xtl = 0.8 * random.random()
_ytl = 0.8 * random.random()
_bounding_box = etag.BoundingBox.from_coords(
_xtl, _ytl, _xtl + 0.2, _ytl + 0.2)
image_labels.add_object(
etao.DetectedObject(label=_label, bounding_box=_bounding_box))
samples.append(
fo.Sample(
filepath=filepath,
ground_truth=fo.ImageLabels(labels=image_labels),
))
dataset = fo.Dataset()
dataset.add_samples(samples)
return dataset
def to_detected_object(self, name=None):
"""Returns an ``eta.core.objects.DetectedObject`` representation of
this instance.
Args:
name (None): the name of the label field
Returns:
an ``eta.core.objects.DetectedObject``
"""
label = self.label
# pylint: disable=unpacking-non-sequence
tlx, tly, w, h = self.bounding_box
brx = tlx + w
bry = tly + h
bounding_box = etag.BoundingBox.from_coords(tlx, tly, brx, bry)
confidence = self.confidence
# pylint: disable=no-member
attrs = etad.AttributeContainer()
for attr_name, attr in self.attributes.items():
attrs.add(etad.CategoricalAttribute(attr_name, attr.value))
return etao.DetectedObject(
label=label,
bounding_box=bounding_box,
confidence=confidence,
name=name,
attrs=attrs,
)
def _make_attribute(name, value):
if isinstance(value, bool):
return etad.BooleanAttribute(name, value)
if etau.is_numeric(value):
return etad.NumericAttribute(name, value)
return etad.CategoricalAttribute(name, value)
def to_eta_attribute(self):
"""Returns an ``eta.core.data.Attribute`` representation of the
attribute.
Returns:
an ``eta.core.data.Attribute``
"""
return etad.CategoricalAttribute(self.name, self.value)
def to_eta_attribute(self):
"""Returns an ``eta.core.data.Attribute`` representation of the
attribute.
Returns:
an ``eta.core.data.Attribute``
"""
if isinstance(self.value, bool):
return etad.BooleanAttribute(self.name, self.value)
if etau.is_numeric(self.value):
return etad.NumericAttribute(self.name, self.value)
return etad.CategoricalAttribute(self.name, self.value)
def _to_eta_attributes(attributes):
attrs = etad.AttributeContainer()
for attr_name, attr in attributes.items():
attr_value = attr.value
if isinstance(attr_value, bool):
_attr = etad.BooleanAttribute(attr_name, attr_value)
elif etau.is_numeric(attr_value):
_attr = etad.NumericAttribute(attr_name, attr_value)
else:
_attr = etad.CategoricalAttribute(attr_name, str(attr_value))
attrs.add(_attr)
return attrs
def to_image_labels(self, name=None):
"""Returns an ``eta.core.image.ImageLabels`` representation of this
instance.
Args:
name (None): the name of the label field
Returns:
an ``eta.core.image.ImageLabels``
"""
image_labels = etai.ImageLabels()
image_labels.add_attribute(
etad.CategoricalAttribute(name,
self.label,
confidence=self.confidence))
return image_labels
def to_schema(self):
"""Returns an ``eta.core.image.ImageLabelsSchema`` representation of
the task labels.
Returns:
an ``eta.core.image.ImageLabelsSchema``
"""
schema = etai.ImageLabelsSchema()
for label in self.labels:
_label = label["name"]
schema.add_object_label(_label)
for attribute in label.get("attributes", []):
_name = attribute["name"]
_categories = attribute["categories"]
for _value in _categories:
_attr = etad.CategoricalAttribute(_name, _value)
schema.add_object_attribute(_label, _attr)
return schema
def to_image_labels(self, name=None):
"""Returns an ``eta.core.image.ImageLabels`` representation of this
instance.
Args:
name (None): the name of the label field
Returns:
an ``eta.core.image.ImageLabels``
"""
image_labels = etai.ImageLabels()
# pylint: disable=not-an-iterable
for classification in self.classifications:
image_labels.add_attribute(
etad.CategoricalAttribute(
name,
classification.label,
confidence=classification.confidence,
))
return image_labels