Enhance the anomaly detection task

The current anomaly detection should be renamed to "Visual Anomaly Detection" I guess..
The most common regime is Single Class detection on MVTec AD, VisA, MVTec AD 2.
Some upcoming regimes are Few-Shot training on these datasets.
The most common metrics are Image AUROC, Segmentation AUROC, Image F1-Max, Segmentation F1 Max.

You find a good overview of historically important methods and some newer methods here: https://anomalib.readthedocs.io/en/lib-v2.6.2/markdown/guides/reference/models/image/index.html

Is there any chance to help with adding the data directly? It's a massive task for one person alone.

No repository files or tests are named. Start by reviewing the existing anomaly detection task and the linked anomalib model overview, then define the scope for the rename and the dataset, regime, metric, and method data. Done means the agreed Visual Anomaly Detection task contains the selected information consistently.

Written by the indexing model from the issue text.