robokudo.utils.hypothesis_comparators

Attributes

THyothesis

Classes

HypothesisComparator

A comparator that can compare two hypotheses.

ObjectHypothesisComparator

Module Contents

robokudo.utils.hypothesis_comparators.THyothesis
class robokudo.utils.hypothesis_comparators.HypothesisComparator(cas_fn: typing_extensions.Callable[[], robokudo.cas.CAS])

Bases: typing_extensions.Generic[THyothesis]

A comparator that can compare two hypotheses.

cas_fn

Function to get a CAS instance, to be used for annotation filtering.

annotation_comparators: typing_extensions.Dict[typing_extensions.Type[robokudo.types.core.Annotation], robokudo.utils.comparators.FeatureComparator]

Feature comparators used by this hypothesis comparator to compare annotations.

type_comparators: typing_extensions.Dict[typing_extensions.Type[robokudo.types.core.Type], robokudo.utils.comparators.FeatureComparator]

Feature comparators used by this hypothesis comparator to compare other robokudo types.

abstract compute_similarity(hypothesis1: THyothesis, hypothesis2: THyothesis) → float

Compute the similarity between two hypotheses.

Parameters:

  • hypothesis1 – The first hypothesis to compare.

  • hypothesis2 – The second hypothesis to compare.

Returns:

The similarity of the two hypotheses between 0.0 (no similarity) and 1.0 (identical).

with_comparator_for(compared_type: typing_extensions.Type[typing_extensions.Union[robokudo.types.core.Annotation, robokudo.types.core.Type]], weight: float, **kwargs) → HypothesisComparator

Add a feature comparator for the given annotation using the given weight to the comparator.

Parameters:

  • compared_type – The type that should be compared by the hypothesis comparator.

  • weight – The weight the hypothesis comparator should use for the compared type.

Returns:

The hypothesis comparator with the added comparator.

Raises:

  • KeyError – If no comparator can be found for the given type.

  • ValueError – If the given type cannot be compared.

class robokudo.utils.hypothesis_comparators.ObjectHypothesisComparator

Bases: HypothesisComparator[robokudo.types.scene.ObjectHypothesis]

compute_similarity(hypothesis1: robokudo.types.scene.ObjectHypothesis, hypothesis2: robokudo.types.scene.ObjectHypothesis) → float

Compute the similarity between two hypotheses.

Parameters:

  • hypothesis1 – The first hypothesis to compare.

  • hypothesis2 – The second hypothesis to compare.

Returns:

The similarity of the two hypotheses between 0.0 (no similarity) and 1.0 (identical).