Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Hierarchical Labeling #742

Merged
merged 3 commits into from
Oct 31, 2022
Merged

Hierarchical Labeling #742

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
acea3b6
hierarchical labeling
bonhunko Oct 19, 2022
3c2795f
add empty line to the end of json/remove duplicated test
bonhunko Oct 20, 2022
1897b0c
Merge branch 'develop' into feature/hierarchical_labeling
wonjuleee Oct 31, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions CHANGELOG.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,6 +13,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
- Add Visualization Python API
- Bbox
(<https://github.com/openvinotoolkit/datumaro/pull/744>)
- Support for exclusive of labels with LabelGroup
(<https://github.com/openvinotoolkit/datumaro/pull/742>)

### Changed
- N/A
Expand Down
24 changes: 23 additions & 1 deletion datumaro/components/annotation.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -98,7 +98,14 @@ class Category:
parent: str = field(default="", validator=default_if_none(str))
attributes: Set[str] = field(factory=set, validator=default_if_none(set))

@attrs(slots=True, order=False)
class LabelGroup:
name: str = field(converter=str, validator=not_empty)
labels: List[str] = field(default=[], validator=default_if_none(list))
group_type: str = field(default="exclusive", validator=default_if_none(str))

items: List[str] = field(factory=list, validator=default_if_none(list))
label_groups: List[str] = field(factory=list, validator=default_if_none(list))
_indices: Dict[str, int] = field(factory=dict, init=False, eq=False)

@classmethod
Expand Down Expand Up @@ -146,7 +153,10 @@ def _reindex(self):
self._indices = indices

def add(
self, name: str, parent: Optional[str] = None, attributes: Optional[Set[str]] = None
self,
name: str,
parent: Optional[str] = None,
attributes: Optional[Set[str]] = None,
) -> int:
assert name
assert name not in self._indices, name
Expand All @@ -156,6 +166,18 @@ def add(
self._indices[name] = index
return index

def add_label_group(
self,
name: str,
labels: List[str],
group_type: str,
) -> int:
assert name

index = len(self.label_groups)
self.label_groups.append(self.LabelGroup(name, labels, group_type))
return index

def find(self, name: str) -> Tuple[Optional[int], Optional[Category]]:
index = self._indices.get(name)
if index is not None:
Expand Down
12 changes: 12 additions & 0 deletions datumaro/plugins/datumaro_format/converter.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -268,9 +268,13 @@ def _convert_cuboid_3d_object(self, obj):
def _convert_attribute_categories(self, attributes):
return sorted(attributes)

def _convert_labels_label_groups(self, labels):
return sorted(labels)

def _convert_label_categories(self, obj):
converted = {
"labels": [],
"label_groups": [],
"attributes": self._convert_attribute_categories(obj.attributes),
}
for label in obj.items:
Expand All @@ -281,6 +285,14 @@ def _convert_label_categories(self, obj):
"attributes": self._convert_attribute_categories(label.attributes),
}
)
for label_group in obj.label_groups:
converted["label_groups"].append(
{
"name": cast(label_group.name, str),
"group_type": cast(label_group.group_type, str),
"labels": self._convert_labels_label_groups(label_group.labels),
}
)
return converted

def _convert_mask_categories(self, obj):
Expand Down
9 changes: 8 additions & 1 deletion datumaro/plugins/datumaro_format/extractor.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -65,7 +65,14 @@ def _load_categories(parsed):
label_categories = LabelCategories(attributes=parsed_label_cat.get("attributes", []))
for item in parsed_label_cat["labels"]:
label_categories.add(
item["name"], parent=item["parent"], attributes=item.get("attributes", [])
item["name"],
parent=item["parent"],
attributes=item.get("attributes", []),
)

for item in parsed_label_cat.get("label_groups", []):
label_categories.add_label_group(
item["name"], labels=item["labels"], group_type=item["group_type"]
)

categories[AnnotationType.label] = label_categories
Expand Down
77 changes: 77 additions & 0 deletions tests/assets/datumaro_dataset/annotations/test.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,77 @@
{
"info": {},
"categories": {
"label": {
"label_groups": [
{
"name": "manmade",
"group_type": "exclusive",
"labels": ["car", "bicycle"]
},
{
"name": "empty_group",
"group_type": "empty",
"labels": ["tom", "mary"]
}
],
"labels": [
{
"name": "car",
"parent": "",
"attributes": []
},
{
"name": "bicycle",
"parent": "",
"attributes": []
},
{
"name": "tom",
"parent": "",
"attributes": []
},
{
"name": "mary",
"parent": "",
"attributes": []
}
],
"attributes": []
}
},
"items": [
{
"id": "c",
"annotations": [
{
"id": 0,
"type": "label",
"attributes": {
"score": 1.0
},
"group": 0,
"label_id": 1
},
{
"id": 0,
"type": "label",
"attributes": {
"score": 1.0
},
"group": 0,
"label_id": 3
}
],
"image": {
"path": "../tests/assets/datumaro_dataset/images/test/c.jpg",
"size": [
10,
5
]
},
"media": {
"path": "../tests/assets/datumaro_dataset/images/test/c.jpg"
}
}
]
}
103 changes: 103 additions & 0 deletions tests/assets/datumaro_dataset/annotations/train.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,103 @@
{
"info": {},
"categories": {
"label": {
"label_groups": [
{
"name": "manmade",
"group_type": "exclusive",
"labels": ["car", "bicycle"]
},
{
"name": "empty_group",
"group_type": "empty",
"labels": ["tom", "mary"]
}
],
"labels": [
{
"name": "car",
"parent": "",
"attributes": []
},
{
"name": "bicycle",
"parent": "",
"attributes": []
},
{
"name": "tom",
"parent": "",
"attributes": []
},
{
"name": "mary",
"parent": "",
"attributes": []
}
],
"attributes": []
}
},
"items": [
{
"id": "a",
"annotations": [
{
"id": 0,
"type": "label",
"attributes": {
"score": 1.0
},
"group": 0,
"label_id": 0
}
],
"image": {
"path": "../tests/assets/datumaro_dataset/images/train/a.jpg"
},
"media": {
"path": "../tests/assets/datumaro_dataset/images/train/a.jpg"
}
},
{
"id": "b",
"annotations": [
{
"id": 0,
"type": "label",
"group": 0,
"label_id": 0
},
{
"id": 1,
"type": "label",
"group": 0,
"label_id": 1
},
{
"id": 2,
"type": "label",
"group": 0,
"label_id": 2
},
{
"id": 3,
"type": "label",
"group": 0,
"label_id": 5
}
],
"image": {
"path": "../tests/assets/datumaro_dataset/images/train/b.jpg",
"size": [
2,
8
]
},
"media": {
"path": "../tests/assets/datumaro_dataset/images/train/b.jpg"
}
}
]
}
54 changes: 54 additions & 0 deletions tests/assets/datumaro_dataset/annotations/validation.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,54 @@
{
"info": {},
"categories": {
"label": {
"label_groups": [
{
"name": "manmade",
"group_type": "exclusive",
"labels": ["car", "bicycle"]
},
{
"name": "empty_group",
"group_type": "empty",
"labels": ["tom", "mary"]
}
],
"labels": [
{
"name": "car",
"parent": "",
"attributes": []
},
{
"name": "bicycle",
"parent": "",
"attributes": []
},
{
"name": "tom",
"parent": "",
"attributes": []
},
{
"name": "mary",
"parent": "",
"attributes": []
}
],
"attributes": []
}
},
"items": [
{
"id": "d",
"annotations": [],
"image": {
"path": "../tests/assets/datumaro_dataset/images/validation/d.png"
},
"media": {
"path": "../tests/assets/datumaro_dataset/images/validation/d.png"
}
}
]
}
Binary file added tests/assets/datumaro_dataset/images/test/c.jpg
View file
Open in desktop
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added tests/assets/datumaro_dataset/images/train/a.jpg
View file
Open in desktop
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added tests/assets/datumaro_dataset/images/train/b.jpg
View file
Open in desktop
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added tests/assets/datumaro_dataset/images/validation/d.png
View file
Open in desktop
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
53 changes: 53 additions & 0 deletions tests/test_labeling.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,53 @@
# Copyright (C) 2019-2022 Intel Corporation
#
# SPDX-License-Identifier: MIT
import tempfile
from unittest.case import TestCase

import numpy as np

from datumaro.components.annotation import AnnotationType, Label, LabelCategories
from datumaro.components.extractor import DatasetItem
from datumaro.components.media import Image
from datumaro.components.project import Dataset

from .requirements import Requirements, mark_requirement


class LabelingTest(TestCase):
@mark_requirement(Requirements.DATUM_GENERAL_REQ)
def test_label_group(self):
label_categories = LabelCategories()
label_categories.add("car", parent="")
label_categories.add("bicycle", parent="")

label_categories.add_label_group("manmade", ["car", "bicycle"], group_type="exclusive")

dataset = Dataset.from_iterable(
[
DatasetItem(
id=0,
subset="train",
media=Image(data=np.ones((10, 6, 3))),
annotations=[
Label(
0,
id=0,
),
Label(
1,
id=1,
),
],
),
],
categories={
AnnotationType.label: label_categories,
},
)

with tempfile.TemporaryDirectory() as temp_home:
dataset.export(temp_home, format="datumaro")
dataset_imported = Dataset.import_from(temp_home, format="datumaro")

self.assertEqual(len(dataset_imported.categories()[AnnotationType.label].label_groups), 1)