CutPaste: Self-Supervised Learning for Anomaly Detection and Localization

TL;DR

This paper introduces CutPaste, a self-supervised learning framework for anomaly detection and localization that achieves state-of-the-art results by learning from normal data only, using a simple data augmentation strategy.

Contribution

The paper proposes a novel two-stage framework utilizing self-supervised representations and a generative classifier, improving defect detection and localization without requiring anomalous training data.

Findings

Achieves 3.1 AUC improvement over previous methods when training from scratch.

Sets a new state-of-the-art 96.6 AUC with transfer learning on ImageNet.

Extends framework for localizing defective areas without annotations.

Abstract

We aim at constructing a high performance model for defect detection that detects unknown anomalous patterns of an image without anomalous data. To this end, we propose a two-stage framework for building anomaly detectors using normal training data only. We first learn self-supervised deep representations and then build a generative one-class classifier on learned representations. We learn representations by classifying normal data from the CutPaste, a simple data augmentation strategy that cuts an image patch and pastes at a random location of a large image. Our empirical study on MVTec anomaly detection dataset demonstrates the proposed algorithm is general to be able to detect various types of real-world defects. We bring the improvement upon previous arts by 3.1 AUCs when learning representations from scratch. By transfer learning on pretrained representations on ImageNet, we achieve a new state-of-theart 96.6 AUC. Lastly, we extend the framework to learn and extract representations from patches to allow localizing defective areas without annotations during training.