Perceptual Image Anomaly Detection

TL;DR

This paper introduces a new perceptual anomaly detection method using GANs, a novel similarity metric, and a hyperparameter tuning approach, achieving state-of-the-art results on image benchmarks.

Contribution

It presents a novel GAN-based framework with a perceptual similarity metric and a new loss weight selection method for anomaly detection.

Findings

Achieves state-of-the-art performance on image anomaly detection benchmarks.

Introduces a robust perceptual similarity metric for abnormality detection.

Develops a hyperparameter tuning method without validation datasets.

Abstract

We present a novel method for image anomaly detection, where algorithms that use samples drawn from some distribution of "normal" data, aim to detect out-of-distribution (abnormal) samples. Our approach includes a combination of encoder and generator for mapping an image distribution to a predefined latent distribution and vice versa. It leverages Generative Adversarial Networks to learn these data distributions and uses perceptual loss for the detection of image abnormality. To accomplish this goal, we introduce a new similarity metric, which expresses the perceived similarity between images and is robust to changes in image contrast. Secondly, we introduce a novel approach for the selection of weights of a multi-objective loss function (image reconstruction and distribution mapping) in the absence of a validation dataset for hyperparameter tuning. After training, our model measures the abnormality of the input image as the perceptual dissimilarity between it and the closest generated image of the modeled data distribution. The proposed approach is extensively evaluated on several publicly available image benchmarks and achieves state-of-the-art performance.