Modeling the Distribution of Normal Data in Pre-Trained Deep Features for Anomaly Detection in PyTorch

PyTorch implementation of Modeling the Distribution of Normal Data in Pre-Trained Deep Features for Anomaly Detection.

This paper presents an anomaly detection approach that consists of fitting a multivariate Gaussian to normal data in the pre-trained deep feature representations, using Mahalanobis distance as anomaly score.
It is a simple yet effective approach and achieves SOTA on MVTec AD dataset.

Prerequisites

• python 3.6+
• PyTorch 1.5+
• efficientnet_pytorch == 0.6.3
• sklearn, matplotlib

Install prerequisites with:

pip install -r requirements.txt

If you already download MVTec AD dataset, move a file to data/mvtec_anomaly_detection.tar.xz.
If you don't have a dataset file, it will be automatically downloaded during the code running.

Usage

To test this implementation code on MVTec AD dataset:

cd src
python main.py

After running the code above, you can see the ROCAUC results in src/result/roc_curve_{model_name}.png

Results

Below is the implementation result of the test set ROCAUC on the MVTec AD dataset.

	Paper	Implementation
bottle	-	100.0
cable	-	94.2
capsule	-	92.3
carpet	-	98.1
grid	-	94.6
hazelnut	-	98.6
leather	-	100.0
metal_nut	-	94.3
pill	-	83.4
screw	-	78.1
tile	-	98.6
toothbrush	-	96.7
transistor	-	96.1
wood	-	98.5
zipper	-	97.7
Average	94.8	94.7

ROC Curve