Motion

Run specification for human motion.

pydantic model moveai_mocap.models.motion.runs.mocap.MocapRun

Run specification for human motion.

Fields:
field actor_mapping: list[ActorMapping] | None = None

Define what actors exist within this capture, along with their skeletons. Bounding boxes define where the actors physical position is at the first frame of the associated video capture input. If provided, will provide more accurate outputs.

Validated by:
field number_of_actors: Annotated[int, Gt(gt=0)] [Required]

How many actors are in this capture.

Constraints:

  • gt = 0

Validated by:
field options: Options | None [Optional]

Configurable options for processing.

Validated by:
field outputs: list[Resource] [Required]

What output resources should be created?

Validated by:
field rig: Resource | None = None

The configuration for the rig to be used for this capture. This is optional and will be excluded from the run configuration if not provided. If not provided, the default rig will be used.

Validated by:
field volume_service: Service | None = None

The source of the capture volume (i.e. S3, GCP etc). Not required if there is only a single video input

Validated by:
field volume_uri_prefix: str | None = None

The path to where the volume files are stored - without any filename. Not required if there is only a single video input

Validated by:
static calculate_iou(bounding_box1: Tuple[float, float, float, float], bounding_box2: Tuple[float, float, float, float]) → float

Calculate the Intersection over Union (IoU) of two bounding boxes.

x1, y1 is the top left corner x2, y2 is the bottom right corner

Args:

bounding_box1: Bounding box 1 list of 4 elements of the form [x1, y1, x2, y2] bounding_box2: Bounding box 2 list of 4 elements of the form [x1, y1, x2, y2]

Returns:

IoU as a float, the ratio of intersection area over union area of the two boxes.

classmethod check_tolerance(device_id: str, pairs: list[Tuple[Tuple[float, float, float, float], Tuple[float, float, float, float]]]) → None

Check the tolerance on the list of pairs of bounding boxes.

Args:

device_id: The device ID to check the bounding boxes for. pairs: The list of pairs of bounding boxes to check.

Raises:

ValueError: If the bounding boxes for a pair are outside tolerance.

validator validate_actor_device_id_on_capture  »  all fields

Validate that the device ids on all actors exist on the Capture.

Raises:

ValueError: At least one device_id on the ActorMapping doesn’t exist on the Capture

Returns:

The validated object

validator validate_actor_mapping  »  all fields

Validate that the number of actors does not exist one if there is a single video input.

Raises:

ValueError: The number of actors is incorrect

Returns:

The validated data

validator validate_number_of_actors  »  all fields

Validate that the a singlecam capture does not have actor mapping.

Raises:

ValueError: The number of actors is incorrect

Returns:

The validated data

validator validate_outputs  »  outputs

Validate that the output URIs are unique.

Args:

outputs: The list of outputs to validate

Raises:

ValueError: There are duplicate resource output uris.

Returns:

The validated data

validator validate_unique_actor  »  actor_mapping

Validate that the actor is unique on the actor mapping.

Uses the actor name to ensure uniqueness

Args:

actor_mapping_value: The actor mapping to validate

Raises:

ValueError: If the actor name is not unique

Returns:

The validated model data

validator validate_volume_uri_prefix  »  all fields

Validate that the input volume is provided if there are multiple video inputs.

Raises:

ValueError: If there are multiple video inputs and no input volume is provided.

Returns:

The validated data

validator validator_bounding_box_overlap  »  actor_mapping

Validate that the bounding box overlap is within tolerances.

The bounding box overlap must be within tolerance for each actor on a capture input.

Args:

actor_mapping_value: The actor mapping to validate

Returns:

The validated model data

Processing options for motion capture data.

pydantic model moveai_mocap.models.motion.options.Options

Processing options for motion capture data.

Fields:
field floor_plane: bool | None = False

Should the floor plane be estimated?

field mocap_model: MocapModel | None = None

The mocap model to use.

field track_ball: bool | None = False

Is there a ball that needs tracking?

field track_fingers: bool | None = False

Should the fingers be tracked?

field track_jersey_numbers: bool | None = False

Should actors be tracked via a jersey number?