# Robust Anomaly Detection in Images using Adversarial Autoencoders

## TL;DR

This paper introduces an adversarial autoencoder-based method for anomaly detection in images that remains effective even when training data contains outliers, improving robustness over traditional autoencoders.

## Contribution

The authors adapt adversarial autoencoders to impose a prior on latent space, enabling early anomaly rejection and robustness against contaminated training sets.

## Key findings

- Adversarial autoencoders improve anomaly detection robustness.
- Training autoencoders on contaminated data reduces outlier detection performance.
- Likelihood-based anomaly rejection enhances detection accuracy.

## Abstract

Reliably detecting anomalies in a given set of images is a task of high practical relevance for visual quality inspection, surveillance, or medical image analysis. Autoencoder neural networks learn to reconstruct normal images, and hence can classify those images as anomalies, where the reconstruction error exceeds some threshold. Here we analyze a fundamental problem of this approach when the training set is contaminated with a small fraction of outliers. We find that continued training of autoencoders inevitably reduces the reconstruction error of outliers, and hence degrades the anomaly detection performance. In order to counteract this effect, an adversarial autoencoder architecture is adapted, which imposes a prior distribution on the latent representation, typically placing anomalies into low likelihood-regions. Utilizing the likelihood model, potential anomalies can be identified and rejected already during training, which results in an anomaly detector that is significantly more robust to the presence of outliers during training.