Energy-Based Anomaly Detection and Localization

TL;DR

This paper proposes a unified energy-based approach for semi-supervised visual anomaly detection and localization, utilizing density estimates and gradient maps to identify and localize anomalies in images, with promising initial results on industrial datasets.

Contribution

It introduces a novel energy-based model framework that jointly detects and localizes anomalies using only anomaly-free training data.

Findings

Energy scores effectively discriminate normal and anomalous images.

Gradient maps provide accurate pixel-level localization of anomalies.

Simple processing of gradient maps can enhance detection performance.

Abstract

This brief sketches initial progress towards a unified energy-based solution for the semi-supervised visual anomaly detection and localization problem. In this setup, we have access to only anomaly-free training data and want to detect and identify anomalies of an arbitrary nature on test data. We employ the density estimates from the energy-based model (EBM) as normalcy scores that can be used to discriminate normal images from anomalous ones. Further, we back-propagate the gradients of the energy score with respect to the image in order to generate a gradient map that provides pixel-level spatial localization of the anomalies in the image. In addition to the spatial localization, we show that simple processing of the gradient map can also provide alternative normalcy scores that either match or surpass the detection performance obtained with the energy value. To quantitatively validate the performance of the proposed method, we conduct experiments on the MVTec industrial dataset. Though still preliminary, our results are very promising and reveal the potential of EBMs for simultaneously detecting and localizing unforeseen anomalies in images.