Augmentation

class maite.protocols.multiobject_tracking.Augmentation(*args, **kwargs)

An augmentation protocol for the multi-object tracking AI problem.

An augmentation is expected to take a batch of data and return a modified version of that batch. Implementers must provide a single method that takes and returns a labeled data batch, where a labeled data batch is represented by a tuple of types Sequence [VideoStream], Sequence [MultiobjectTrackingTarget], and Sequence [DatumMetadata]. These correspond to the model input batch type, model target batch type, and datum-level metadata batch type, respectively.

Attributes:

metadataAugmentationMetadata

A typed dictionary containing at least an ‘id’ field of type str

Methods

__call__(datum: tuple[Sequence[VideoStream], Sequence[MultiobjectTrackingTarget], Sequence[DatumMetadata]]) -> tuple[Sequence[VideoStream], Sequence[MultiobjectTrackingTarget], Sequence[DatumMetadata]])

Return a modified version of original data batch. A data batch is represented by a tuple of model input batch (as Sequence [VideoStream], model target batch (as Sequence [MultiobjectTrackingTarget], and batch metadata (as Sequence [DatumMetadata]), respectively.