GANomaly: Semi-supervised Anomaly Detection via Adversarial Training

Abstract

Anomaly detection is a classical problem in computer vision, namely the determination of the normal from the abnormal when datasets are highly biased towards one class (normal) due to the insufficient sample size of the other class (abnormal). While this can be addressed as a supervised learning problem, a significantly more challenging problem is that of detecting the unknown/unseen anomaly case that takes us instead into the space of a one-class, semi-supervised learning paradigm. We introduce such a novel anomaly detection model, by using a conditional generative adversarial network that jointly learns the generation of high-dimensional image space and the inference of latent space. Employing encoder-decoder-encoder sub-networks in the generator network enables the model to map the input image to a lower dimension vector, which is then used to reconstruct the generated output image. The use of the additional encoder network maps this generated image to its latent representation. Minimizing the distance between these images and the latent vectors during training aids in learning the data distribution for the normal samples. As a result, a larger distance metric from this learned data distribution at inference time is indicative of an outlier from that distribution -an anomaly. Experimentation over several benchmark datasets, from varying domains, shows the model efficacy and superiority over previous state-of-the-art approaches.

MimicGAN: Robust Projection onto Image Manifolds with Corruption Mimicking

“…Area under ROC VAE [7] 0.34±0.18 AnoGAN [7] 0.44±0.07 GANomaly (BIGAN) [6] 0.78 ± 0.11 EGBAD (PGD baseline) [58] 0.50±0.13 MimicGAN (ours) 0.78 ± 0.14 Table 2: Anomaly Detection on MNIST leave-one-class out experiment. Average performance on all 10 classes is reported below.…”

“…In this section, we study how robust projections can lead to performance gains in different applications involving GANs. The idea of projecting onto a known image manifold has been leveraged in applications such as adversarial defense [48,27,50], anomaly detection [58,6,7], domain adaptation, etc. We use the following experiments to verify our hypothesis that a more robust projection can make GAN based solutions more effective in these applications.…”

“…GANs have become a popular choice for unsupervised anomaly detection since out of distribution samples are represented as those samples with a relatively high reprojection error, since they are not well represented by the image manifold inferred by the generator. A typical experimental setup [58,6,7] for a k class dataset is to train a GAN on k − 1 classes, and use the k th class as the anomalythis is repeated for every class. We test the effectiveness of MimicGAN in such a task, which we expect to improve with more robust projections.…”

“…We test the effectiveness of MimicGAN in such a task, which we expect to improve with more robust projections. We also compare our method with recent manifold-projection based anomaly detection techniques [7,6,58].…”

“…As a result, they have been effective in several applications ranging from super resolution [35] and image editing [44,57], to image-to-image translation [60,37], etc. Additionally, an accurate parameterization of the image manifold provides a powerful regularization -sometimes referred as a generative prior -for a range of problems such as defense against adversarial attacks [48,27], anomaly detection [6,58], and compressive sensing [52]. Despite the variabilities in their goals and formulations, an overarching requirement in all these applications is the ability to project an image observation onto the image manifold at test time.…”