SemanticSegmentationEvaluator#

class SemanticSegmentationEvaluator[source]#

Bases: ClassificationEvaluator

Evaluates predictions for a set of scenes.

__init__(class_config: ClassConfig, output_uri: str | None = None)#
Parameters:

Methods

__init__(class_config[, output_uri])

create_evaluation()

evaluate_predictions(ground_truth, predictions)

Evaluate predictions against ground truth.

evaluate_scene(scene)

Override to pass null_class_id to filter_by_aoi().

process(scenes[, tmp_dir])

Evaluate all given scenes and save the evaluations.
__init__(class_config: ClassConfig, output_uri: str | None = None)#
Parameters:
create_evaluation() → SemanticSegmentationEvaluation[source]#
Return type:

SemanticSegmentationEvaluation

evaluate_predictions(ground_truth: Labels, predictions: Labels) → ClassificationEvaluation#

Evaluate predictions against ground truth.

Parameters:

  • ground_truth (Labels) – Ground truth labels.

  • predictions (Labels) – Predictions.

Returns:

The evaluation.

Return type:

Any

evaluate_scene(scene: Scene) → SemanticSegmentationEvaluation[source]#

Override to pass null_class_id to filter_by_aoi().

Parameters:

scene (Scene) –

Return type:

SemanticSegmentationEvaluation

process(scenes: Iterable[Scene], tmp_dir: str | None = None) → None#

Evaluate all given scenes and save the evaluations.

Parameters:

  • scenes (Iterable[Scene]) – Scenes to evaluate.

  • tmp_dir (str | None) –

Return type:

None