Unsupervised Learning of Anomaly Detection from Contaminated Image Data using Simultaneous Encoder Training

TL;DR

This paper introduces a novel encoder training approach within GANs to improve unsupervised anomaly detection in contaminated image datasets, effectively handling real-world data imperfections and achieving state-of-the-art results.

Contribution

It proposes joint generator-encoder training to mitigate contamination effects in GAN-based anomaly detection, enhancing detection accuracy in real-world scenarios.

Findings

Encoder training improves anomaly detection in contaminated data.

The method achieves state-of-the-art results on CIFAR-10.

Effective in detecting anomalies in cell image datasets.

Abstract

Unsupervised learning of anomaly detection in high-dimensional data, such as images, is a challenging problem recently subject to intense research. Through careful modelling of the data distribution of normal samples, it is possible to detect deviant samples, so called anomalies. Generative Adversarial Networks (GANs) can model the highly complex, high-dimensional data distribution of normal image samples, and have shown to be a suitable approach to the problem. Previously published GAN-based anomaly detection methods often assume that anomaly-free data is available for training. However, this assumption is not valid in most real-life scenarios, a.k.a. in the wild. In this work, we evaluate the effects of anomaly contaminations in the training data on state-of-the-art GAN-based anomaly detection methods. As expected, detection performance deteriorates. To address this performance drop, we propose to add an additional encoder network already at training time and show that joint generator-encoder training stratifies the latent space, mitigating the problem with contaminated data. We show experimentally that the norm of a query image in this stratified latent space becomes a highly significant cue to discriminate anomalies from normal data. The proposed method achieves state-of-the-art performance on CIFAR-10 as well as on a large, previously untested dataset with cell images.