Patch SVDD: Patch-level SVDD for Anomaly Detection and Segmentation

TL;DR

This paper introduces Patch SVDD, a patch-based deep learning extension of SVDD for improved image anomaly detection and segmentation, achieving significant performance gains on the MVTec AD dataset.

Contribution

It extends SVDD into a patch-based, self-supervised learning framework for simultaneous anomaly detection and segmentation, enhancing accuracy over previous methods.

Findings

Detection AUROC increased by 9.8%

Segmentation AUROC increased by 7.0%

Effective for industrial anomaly detection

Abstract

In this paper, we address the problem of image anomaly detection and segmentation. Anomaly detection involves making a binary decision as to whether an input image contains an anomaly, and anomaly segmentation aims to locate the anomaly on the pixel level. Support vector data description (SVDD) is a long-standing algorithm used for an anomaly detection, and we extend its deep learning variant to the patch-based method using self-supervised learning. This extension enables anomaly segmentation and improves detection performance. As a result, anomaly detection and segmentation performances measured in AUROC on MVTec AD dataset increased by 9.8% and 7.0%, respectively, compared to the previous state-of-the-art methods. Our results indicate the efficacy of the proposed method and its potential for industrial application. Detailed analysis of the proposed method offers insights regarding its behavior, and the code is available online.