Gaussian & Mahalanobis distance
just image level detection, no location.

- fitting a multivariate Gaussian (MVG) per feature level. There are D feature levels.
- each gaussian ⇒ one score by Mahalanobis distance.
- final anomaly score := non weighted summation of the above scores.
References
Modeling the Distribution of Normal Data in Pre-Trained Deep Features for Anomaly Detection, ICPR20