paralleldomain.model.annotation.annotation_types

class AnnotationTypes

Type-safe access to annotations and their associated metadata

BoundingBoxes2D - Two-Dimensional bounding box annotations

BoundingBoxes3D - Three-Dimensional bounding box annotations

SemanticSegmentation2D  - Two-Dimensional semantic segmentation annotations

InstanceSegmentation2D  - Two-Dimensional instance segmentation annotations

SemanticSegmentation3D  - Three-Dimensional semantic segmentation annotations

InstanceSegmentation3D  - Three-Dimensional instance segmentation annotations

OpticalFlow - Optical flow mask annotation

BackwardOpticalFlow - Backwards optical flow mask annotation

Depth - Depth mask annotation

SurfaceNormals2D - Two-Dimensional surface normal annotation

SurfaceNormals3D - Three-Dimensional surface normal annotation

SceneFlow - Point-Cloud scene flow annotation

BackwardSceneFlow - Point-Cloud backwards scene flow annotation

MaterialProperties2D - Two-Dimensional material properties annotation

MaterialProperties3D - Three-Dimensional material properties annotation

Albedo2D - Two-Dimensional albedo documentation

Points2D - Two-Dimensional key point annotations

Points3D - Three-Dimensional key point annotations

Polygons2D - Two-Dimensional polygon annotation

Polygons3D - Three-Dimensional polygon annotation

Polylines2D - Two-Dimensional polyline annotation

Polylines3D - Three-Dimensional polyline annotation

PointCaches - Three-Dimensional point cache annotation

Examples

Access 2D Bounding Box annotations for a camera frame:

>>> camera_frame: SensorFrame = ...  # get any camera's SensorFrame
>>> camera_frame: SensorFrame = ...  # get any camera's SensorFrame
>>>
>>> boxes_2d = camera_frame.get_annotations(AnnotationTypes.BoundingBoxes2D)
>>> for b in boxes_2d.boxes:
>>>     print(b.class_id, b.instance_id)

Access class map for an annotation type in a scene:

>>> scene: Scene = ...  # get a Scene instance
>>>
>>> from paralleldomain.model.annotation import AnnotationTypes
>>>
>>> class_map = scene.get_class_map(AnnotationTypes.SemanticSegmentation2D)
>>> for id, class_detail in class_map.items():
>>>     print(id, class_detail.name)