Efficient GAN-Based Anomaly Detection

About

Generative adversarial networks (GANs) are able to model the complex highdimensional distributions of real-world data, which suggests they could be effective for anomaly detection. However, few works have explored the use of GANs for the anomaly detection task. We leverage recently developed GAN models for anomaly detection, and achieve state-of-the-art performance on image and network intrusion datasets, while being several hundred-fold faster at test time than the only published GAN-based method.

Houssam Zenati, Chuan Sheng Foo, Bruno Lecouat, Gaurav Manek, Vijay Ramaseshan Chandrasekhar · 2018

Related benchmarks

Task	Dataset	Result
Anomaly Detection	MVTecAD (test)	Bottle Score68	55
Anomaly Detection	MVTec AD	AUROC0.6	35
Anomaly Detection	MVTec Anomaly Detection 1.0 (test)	PRO (Carpet)0.52	27
Unsupervised Anomaly Detection	MNIST (Heldout Digit 1)	AUPRC28.7	16
Unsupervised Anomaly Detection	MNIST Heldout Digit 4	AUPRC44.3	16
Unsupervised Anomaly Detection	MNIST Heldout Digit 5	AUPRC51.4	16
Unsupervised Anomaly Detection	MNIST Heldout Digit 7	AUPRC0.347	16
Unsupervised Anomaly Detection	MNIST Heldout Digit 9	AUPRC30.7	16
Anomaly Detection	MNIST	AUPRC (Digit 1)0.287	7
Anomaly Detection	MNIST Heldout Digit 1 (test)	AUPRC28.7	7