Is it possible to do multi-class target detection?

[lucianchauvin]

Is there an example anywhere, or is it even possible, to do multi-class target detection using anomaly with the openvino inference? Say, I train my model, and having anomaly detection is great but can I now identify which anomaly is being detected if I have multiple targets? Thanks.

[abc-125]

As far as I know, anomaly detection models in anomalib don't have such an ability. They calculate just a score, which shows how normal or abnormal the image is.