anomalib.models.cflow.anomaly_map

Anomaly Map Generator for CFlow model implementation.

Module Contents

Classes

AnomalyMapGenerator

Generate Anomaly Heatmap.
class anomalib.models.cflow.anomaly_map.AnomalyMapGenerator(image_size: Union[omegaconf.ListConfig, Tuple], pool_layers: List[str])[source]

Generate Anomaly Heatmap.

compute_anomaly_map(distribution: Union[List[torch.Tensor], List[List]], height: List[int], width: List[int]) → torch.Tensor[source]

Compute the layer map based on likelihood estimation.

Parameters

  • distribution – Probability distribution for each decoder block

  • height – blocks height

  • width – blocks width

Returns

Final Anomaly Map

__call__(**kwargs: Union[List[torch.Tensor], List[int], List[List]]) → torch.Tensor[source]

Returns anomaly_map.

Expects distribution, height and ‘width’ keywords to be passed explicitly

Example >>> anomaly_map_generator = AnomalyMapGenerator(image_size=tuple(hparams.model.input_size), >>> pool_layers=pool_layers) >>> output = self.anomaly_map_generator(distribution=dist, height=height, width=width)

Raises

ValueError – distribution, height and ‘width’ keys are not found

Returns

anomaly map

Return type

torch.Tensor