mednet.engine.saliency.evaluator

Functions

run(saliency_map_algorithm, completeness, ...)	Evaluate multiple saliency map algorithms and produces summarized results.
summary_table(summary, fmt)	Tabulate various summaries into one table.

mednet.engine.saliency.evaluator.summary_table(summary, fmt)

Tabulate various summaries into one table.

Parameters:

• summary (dict[Literal['ablationcam', 'eigencam', 'eigengradcam', 'fullgrad', 'gradcam', 'gradcamelementwise', 'gradcam++', 'gradcamplusplus', 'hirescam', 'layercam', 'randomcam', 'scorecam', 'xgradcam'], dict[str, Any]]) – A dictionary mapping saliency algorithm names to the results of run().

• fmt (str) – One of the formats supported by python-tabulate.

Returns:

A string containing the tabulated information.

Return type:

str

mednet.engine.saliency.evaluator.run(saliency_map_algorithm, completeness, interpretability)

Evaluate multiple saliency map algorithms and produces summarized results.

Parameters:

• saliency_map_algorithm (Literal['ablationcam', 'eigencam', 'eigengradcam', 'fullgrad', 'gradcam', 'gradcamelementwise', 'gradcam++', 'gradcamplusplus', 'hirescam', 'layercam', 'randomcam', 'scorecam', 'xgradcam']) – The algorithm for saliency map estimation that is being analysed.

• completeness (dict[str, list]) – A dictionary mapping dataset names to tables with the sample name and completness (among which Average ROAD) scores.

• interpretability (dict[str, list]) – A dictionary mapping dataset names to tables with the sample name and interpretability (among which Prop. Energy) scores.

Returns:

A dictionary with most important statistical values for the main completeness (AOPC-Combined), interpretability (Prop. Energy), and a combination of both (ROAD-Weighted Prop. Energy) scores.

Return type:

dict[str, Any]