VideoFigure¶

class VideoFigure[source]¶

Bases: object

VideoFigure object for VideoAnnotation. VideoFigure object is immutable.

Parameters

video_object : VideoObject: VideoObject object.
geometry : Geometry: Label geometry.
frame_index : int: Index of Frame to which VideoFigure belongs.
key_id_map : KeyIdMap, optional: KeyIdMap object.
class_id : int, optional: ID of VideoObject to which VideoFigure belongs.
labeler_login : str, optional: Login of the user who created VideoFigure.
updated_at : str, optional: Date and Time when VideoFigure was modified last. Date Format: Year:Month:Day:Hour:Minute:Seconds. Example: ‘2021-01-22T19:37:50.158Z’.
created_at : str, optional: Date and Time when VideoFigure was created. Date Format is the same as in “updated_at” parameter.
track_id : str, optional: ID of the track to which VideoFigure belongs.
smart_tool_input : dict, optional: Smart Tool parameters that were used for labeling.
priority : int, optional: Priority of the figure (position of the figure relative to other overlapping or underlying figures).
status : LabelingStatus, optional: Sets labeling status. Shows how label was created and corrected.

Usage example

import supervisely as sly

obj_class_car = sly.ObjClass('car', sly.Rectangle)
video_obj_car = sly.VideoObject(obj_class_car)
fr_index = 7
geometry = sly.Rectangle(0, 0, 100, 100)
video_figure_car = sly.VideoFigure(video_obj_car, geometry, fr_index)
video_figure_car_json = video_figure_car.to_json()
print(video_figure_car_json)
# Output: {
#     "key": "5e8afd2e26a54ab18154b355fa9665f8",
#     "objectKey": "5860b7a5519b4de7b3d9c1720a40b38a",
#     "geometryType": "rectangle",
#     "geometry": {
#         "points": {
#             "exterior": [
#                 [
#                     0,
#                     0
#                 ],
#                 [
#                     100,
#                     100
#                 ]
#             ],
#             "interior": []
#         }
#     },
#    "nnCreated": false,
#    "nnUpdated": false
# }

Methods

`clone`	Makes a copy of VideoFigure with new fields, if fields are given, otherwise it will use fields of the original VideoFigure.
`from_json`	Convert a json dict to VideoFigure.
`get_meta`	Get metadata for the video figure.
`key`	Figure key.
`to_json`	Convert the VideoFigure to a json dict.
`validate_bounds`	Checks if given image with given size contains a figure.

Attributes

`frame_index`	Frame index of the current VideoFigure.
`geometry`	Geometry of the current VideoFigure.
`parent_object`	VideoObject of current VideoFigure.
`priority`	Priority of the figure (position of the figure relative to other overlapping or underlying figures).
`smart_tool_input`	Smart Tool parameters that were used for labeling.
`status`	Labeling status.
`video_object`	VideoObject of current VideoFigure.

clone(video_object=None, geometry=None, frame_index=None, key=None, class_id=None, labeler_login=None, updated_at=None, created_at=None, track_id=None, smart_tool_input=None, priority=None, status=None)[source]¶

Makes a copy of VideoFigure with new fields, if fields are given, otherwise it will use fields of the original VideoFigure.

Parameters

video_object : VideoObject, optional: VideoObject object.
geometry : Geometry, optional: Label geometry.
frame_index : int, optional: Index of Frame to which VideoFigure belongs.
key_id_map : KeyIdMap, optional: KeyIdMap object.
class_id : int, optional: ID of ObjClass to which VideoFigure belongs.
labeler_login : str, optional: Login of the user who created VideoFigure.
updated_at : str, optional: Date and Time when VideoFigure was modified last. Date Format: Year:Month:Day:Hour:Minute:Seconds. Example: ‘2021-01-22T19:37:50.158Z’.
created_at : str, optional: Date and Time when VideoFigure was created. Date Format is the same as in “updated_at” parameter.
track_id : str, optional: ID of the track to which VideoFigure belongs.
smart_tool_input : dict, optional: Smart Tool parameters that were used for labeling.
priority : int, optional: Priority of the figure (position of the figure relative to other overlapping or underlying figures).
status : LabelingStatus, optional: Sets labeling status. Specifies if the VideoFigure was created by NN model, manually or created by NN and then manually corrected.

Returns

VideoFigure object

Return type

VideoFigure

Usage example

import supervisely as sly

obj_class_car = sly.ObjClass('car', sly.Rectangle)
video_obj_car = sly.VideoObject(obj_class_car)
fr_index = 7
geometry = sly.Rectangle(0, 0, 100, 100)
video_figure_car = sly.VideoFigure(video_obj_car, geometry, fr_index)

obj_class_bus = sly.ObjClass('bus', sly.Rectangle)
video_obj_bus = sly.VideoObject(obj_class_bus)
fr_index_bus = 15
geometry_bus = sly.Rectangle(0, 0, 500, 600)

# Remember that VideoFigure object is immutable, and we need to assign new instance of VideoFigure to a new variable
video_figure_bus = video_figure_car.clone(video_object=video_obj_bus, geometry=geometry_bus, frame_index=fr_index_bus)
print(video_figure_bus.to_json())
# Output: {
#     "key": "c2f501e94f42483ebd202697608e8d26",
#     "objectKey": "942c79137b4547c59193276317f73897",
#     "geometryType": "rectangle",
#     "geometry": {
#         "points": {
#             "exterior": [
#                 [
#                     0,
#                     0
#                 ],
#                 [
#                     600,
#                     500
#                 ]
#             ],
#             "interior": []
#         }
#     }
# }

classmethod from_json(data, objects, frame_index, key_id_map=None)[source]¶

Convert a json dict to VideoFigure. Read more about Supervisely format.

Parameters

data : dict: Dict in json format.
objects : VideoObjectCollection: VideoObjectCollection object.
frame_index : int: Index of Frame to which VideoFigure belongs.
key_id_map : KeyIdMap, optional: KeyIdMap object.

Raises

RuntimeError, if video object ID and video object key are None, if video object key and key_id_map are None, if video object with given id not found in key_id_map

Returns

VideoFigure object

Return type

VideoFigure

Usage example

import supervisely as sly

# Create VideoFigure from json we use data from example to_json(see above)
new_video_figure = sly.VideoFigure.from_json(video_figure_json, sly.VideoObjectCollection([video_obj_car]), fr_index)

get_meta()[source]¶

Get metadata for the video figure.

Returns

Dictionary with metadata for the video figure.

Return type

Dict[str, int]

Usage example

import supervisely as sly

obj_class_car = sly.ObjClass('car', sly.Rectangle)
video_obj_car = sly.VideoObject(obj_class_car)
fr_index = 7
geometry = sly.Rectangle(0, 0, 100, 100)
video_figure_car = sly.VideoFigure(video_obj_car, geometry, fr_index)

print(video_figure_car.get_meta()) # {'frame': 7}

key()[source]¶

Figure key.

Returns

Figure key.

Return type

UUID

Usage example

key = video_figure_car.key
print(key) # 158e6cf4f4ac4c639fc6994aad127c16

to_json(key_id_map=None, save_meta=False)[source]¶

Convert the VideoFigure to a json dict. Read more about Supervisely format.

Parameters

key_id_map : KeyIdMap, optional: KeyIdMap object.
save_meta : bool, optional: Save frame index or not.

Returns

Json format as a dict

Return type

dict

Usage example

import supervisely as sly

obj_class_car = sly.ObjClass('car', sly.Rectangle)
video_obj_car = sly.VideoObject(obj_class_car)
fr_index = 7
geometry = sly.Rectangle(0, 0, 100, 100)
video_figure_car = sly.VideoFigure(video_obj_car, geometry, fr_index)
video_figure_json = video_figure_car.to_json(save_meta=True)
print(video_figure_json)
# Output: {
#     "key": "591d0511ba28462c8cd657691743359c",
#     "objectKey": "e061bc50bd464c23a008b712d195570a",
#     "geometryType": "rectangle",
#     "geometry": {
#         "points": {
#             "exterior": [
#                 [
#                     0,
#                     0
#                 ],
#                 [
#                     100,
#                     100
#                 ]
#             ],
#             "interior": []
#         }
#     },
#     "meta": {"frame": 7},
#     "nnCreated": false,
#     "nnUpdated": false
# }

validate_bounds(img_size, _auto_correct=False)[source]¶

Checks if given image with given size contains a figure.

Parameters

img_size : Tuple[int, int]: Size of the image (height, width).
_auto_correct : bool, optional: Correct the geometry of a shape if it is out of bounds or not.

Raises

OutOfImageBoundsException, if figure is out of image bounds

Returns

None

Return type

NoneType

Usage Example

import supervisely as sly

obj_class_car = sly.ObjClass('car', sly.Rectangle)
video_obj_car = sly.VideoObject(obj_class_car)
fr_index = 7
geometry = sly.Rectangle(0, 0, 100, 100)
video_figure_car = sly.VideoFigure(video_obj_car, geometry, fr_index)

im_size = (50, 200)
video_figure_car.validate_bounds(im_size)
# raise OutOfImageBoundsException("Figure is out of image bounds")

property frame_index¶

Frame index of the current VideoFigure.

Returns

Index of Frame to which VideoFigure belongs

Return type

int

Usage example

fr_index = video_figure_car.frame_index
print(fr_index) # 7

property geometry¶

Geometry of the current VideoFigure.

Returns

Geometry object

Return type

Geometry

Usage example

geometry = video_figure_car.geometry
print(geometry.to_json())
# Output: {
#     "points": {
#         "exterior": [
#             [
#                 0,
#                 0
#             ],
#             [
#                 100,
#                 100
#             ]
#         ],
#         "interior": []
#     }
# }

property parent_object¶

VideoObject of current VideoFigure.

Returns

VideoObject object

Return type

VideoObject

Usage example

video_obj_car = video_figure_car.parent_object
print(video_obj_car.to_json())
# Output: {
#     "key": "d573c6f081544e3da20022d932b259c1",
#     "classTitle": "car",
#     "tags": []
# }

property priority¶: Priority of the figure (position of the figure relative to other overlapping or underlying figures).

property smart_tool_input¶

Smart Tool parameters that were used for labeling.

Example:

{
‘crop’: [[85.69912274538524, 323.07711452375236], [1108.5635719011857, 1543.1199742240174]], ‘visible’: True, ‘negative’: [], ‘positive’: [[597, 933], [474.5072466934964, 1381.6437133813354]]

}

property status¶

Labeling status. Specifies if the VideoFigure was created by NN model, manually or created by NN and then manually corrected.

Return type: LabelingStatus

property video_object¶

VideoObject of current VideoFigure.

Returns

VideoObject object

Return type

VideoObject

Usage example

video_obj_car = video_figure_car.video_object
print(video_obj_car.to_json())
# Output: {
#     "key": "d573c6f081544e3da20022d932b259c1",
#     "classTitle": "car",
#     "tags": []
# }