Train an Anomaly Detector for Images

This example shows how to create an anomaly detector for a specific dataset of images.

Obtain a dataset of images.

Train an anomaly detector on the dataset.

Obtain information about the anomaly detector.

Use the detector on new images.

Compare with the results obtained on corrupted images.

Visualize the behavior of the probability to find a rarer image (RarerProbability) for various levels of corruption.